P-

<-*

o

O

CD

m

THE LIFE OF THE SALMON

Young- salmon artificially reared at the Duke of Richmond and Gordon's ponds near the month
of the Spey. From above downwards the fish are one year, two years, and three years old.
The fourth and largest fish is also three years old, but has spent its third year in a sea-water
pond.

THE

LIFE OF THE SALMON

WITH REFERENCE MORE ESPECIALLY
TO THE FISH IN SCOTLAND

BY

W. L. CALDERWOOD, F.R.S.E.

INSPECTOR OF SALMON FISHERIES FOR SCOTLAND

SECOND IMPRESSION

LONDON
EDWARD ARNOLD

1908

[All rights reserved]

TO

THE DISTRICT FISHERY BOARDS
OF SCOTLAND

I RESPECTFULLY DEDICATE
THIS BOOK

CONTENTS

INTRODUCTORY . . . . . . . . Page xiii

CHAPTER I
Smolts

Early ideas concerning smolts Rearing of smolts in captivity
Marine versus fresh water origin of the salmon
Hatching fry from smolt eggs British smolts descend
when two years old Marking of smolts Habits in migrat-
ing to the sea Autumn migrations Development of
smolts in sea Herr Dahl's capture of smolts in the sea ofi
Norway ' ... Pp. 1-29

CHAPTER II

Grilse

Return of marked smolts as grilse the following season Habits
of grilse in the sea Ascent of grilse from the sea Habit
of many to remain in sea beyond grilse stage Small spring
fish The spawning of grilse Ascent of grilse when in
poor condition Argument in support of the marine origin
of the salmon Entrance of salmon to fresh water not
entirely explained by condition of satiety Reproduction in
fresh water Possible transference of small fish from one
river to another Movements of marked grilse along coast
of Scotland Pp. 80-55

60258

viii CONTENTS

OH AFTER III

Results of Salmon Marking

Early marking operations Modern method of marking
Divided migration Scottish and Irish records showing
short and long periods in the sea Increase of weight in
kelts recaptured as clean fish Significance of divided
migration in regulating fisheries Salmon returning to their
own river Recaptures in other rivers Recaptures on the
coast Direction of movements on coast Kelts recaptured
as kelts Spring fish marked and recaptured Pp. 56-92

CHAPTER IV

The Scales of Salmon as Records of the
Salmon's Life

Age of salmon to be told from their scales Progressive growth
of the scales Lines of growth on scales The spawning
mark Records of marked smolts as shown on scales
Records of small spring fish marked as smolts Infrequency
of the spawning mark Scales of kelt recaptured as clean
fish The average number of lines added Scales of large
salmon Salmon disease The bacillus of origin Fungus
spreads quickly on dead fish .... Pp. 93-109

CHAPTER V

The Feeding of Salmon and Growth of
the Genitalia

Absence of food in salmon taken in fresh water Feeding habits
of fry Food of salmon in the sea Nutriment in tissues of
estuary as against upper water fish Transference of
nutriment from muscles to reproductive organs

Pp. 110-124

CONTENTS

IX

CHAPTER VI
The Salmon and Water Temperature

Early and late rivers Rivers made late by man Temperature
of the sea round Scotland Temperatures of early and late
Scottish rivers Temperature no indication of seasonal
character Effects of temperature on fish ascending rivers
Early runs of fish in upper waters, Garry and Orchy
Effects of cold water upon ascending fish The entrance of
tributaries from main river Rapid ascent of summer fish
Sea lice may remain attached for four or five day&

Pp. 125-145

APPENDIX

A record of salmon marking in Scotland previous to the opera-
tions of the Fishery Board in 1896 . . Pp. 147-157

INDEX

. 159-160

LIST OF ILLUSTRATIONS

Young salmon artificially reared at the Duke
of Richmond and Gordon's ponds near the

mouth of the Spey Frontispiece

PLATE I. Dorsal fin of grilse, caught in 1906,

with silver mark inserted when the fish

was a smolt in 1905 TO face page 30

PLATE II. Measuring length of salmon on

marking board and piercing dorsal fin for

insertion of mark . . . , 58

PLATE III. Weighing fish suspended by worsted

tail band after insertion of mark . . 58
PLATE IV. A male salmon in full spawning

livery, fresh from the sea in November.

The fish has just been marked on the

dorsal fin ......,, 92

PLATE V. Scales of grilse and small spring

salmon recaptured in 1906 and 1907

respectively, each fish having been marked

as a smolt in 1905 * . . . 100

PLATE VI. Two scalesi from a salmon at

different stages of growth, when the fish

was marked as a kelt and recaptured as a

clean fish . . . . 102

PLATE VII. A salmonmarking crew on the

Helmsdale, Sutherland . . . . 122

PREFATORY NOTE

I AM indebted to the Royal Society of Edinburgh for per-
mission to reproduce the Frontispiece and Plate I., which
have already appeared in the Proceedings of the Society
as Illustrations to papers of my own.

I desire, also, to thank Sir Herbert Maxwell for valued
Criticism with regard to certain views I have ventured to
express in this little book, and Mr. H. W. Johnston for
many kindnesses in facilitating the experimental work in
which we have been much associated.

W. L. C.

CARE BRIDGE,

August 15, 1907.

INTRODUCTORY

To all who endeavour to understand the life of the
salmon, two disadvantages are naturally present.
The fish cannot at all stages of its life be brought
under observation ; it goes from us into the ewigkeit
of the ocean, where on the whole it spends more
time than in our rivers, and where its doings excite
men's imaginations and their faculty for romance.
Secondly, the fish belongs to a genus so susceptible
to the influences of environment that it has to be
studied in different localities and under varied con-
ditions, if we would escape a too narrow vision of
its habits and characteristics.

Even then we have to bear in mind that what
we actually see may not always be believed, that
exceptional appearances have to be suspected and
strange occurrences investigated ; while the whole
history of salmon literature reminds us that to write
on this subject is to open up argument and to court
dispute. I have heard of a Cabinet Minister who,
when asked to advance salmon fishery legislation,
promptly declined, saying he wished to die in his bed.
The heat and intolerance which have been exhibited,
in Scotland at least, over disputes about the life of
the salmon almost rival the spirit shown in recent

xiv INTRODUCTORY

times over Church distinctions. Precisely similar
lines of investigation have led contending observers
to completely opposite deductions, the reason being
that the observers were in reality striving for
supremacy rather than for truth.

No wonder, then, that critics find safety in saying
we know very little about the salmon. They are,
even at the present day, perhaps not far wrong.
But writers seem to have acquired the habit of safe-
guarding themselves by emphasising the " mystery "
or " dire perplexity " in which the whole subject is
shrouded. One of the most recent tells us that the
result of the whole body of literature on the salmon
is to give the Philistine " numerous erroneous ideas
on the subject, along with a few ' proved facts.' *
It is clear that he does not think much of the
" proved facts " regards them apparently as on the
same level with the judge's famous characterisation
of the evidence of the skilled witness.

With humility, however, we must plod on search-
ing for our facts if we hope to gain that true con-
ception of our subject which can alone give us a
scientific basis from which to extend a sound regu-
lative treatment of our salmon fisheries. Alexander
Russel has truly said, concerning the importance of
this matter, that " without some knowledge of how,
when, and where the fish breeds, dwells, and feeds,
it is useless to speak and unsafe to act."

The general principle lying at the root of all
salmon fishery legislation has been to counteract
the natural tendency of man to over-fish, to capture
fish at a harmful time or in a harmful way, or to

INTRODUCTORY xv

erect engines which obstructed the ascent of salmon
and facilitated their capture in undue numbers.
The Scottish statutes of four hundred years ago
commonly put down practices which "destroy the
breed of fish, and hurt the commoun profit e of the
realme." From earliest times it has been considered
necessary to protect spawning fish by close seasons,
and to do this satisfactorily observation as to the
breeding season has been inevitable. Like the
crime of sheep-stealing, the crime of poaching
salmon in close time was punishable by death. In
an Act of the first Parliament of James I. (1424) we
find " Quha sa ever be convict of slauchter of
Salmonde in time forbidden be the law, he sail pay
fourtie shillings for the unlaw, and at the third time,
gif he be convict of sik trespasse, he sail tyne his
life, or then bye it." The transference from forty
shillings (Scots) to the death penalty seems rapid,
nor is any indication given as to what may be
regarded as the ultimate price of a poacher's life.
One thing seems clear, however : those early Scottish
legislators were determined to preserve the breed
of salmon. God bless their memory ! In more
recent times we have had not a few examples of
how easy it is to destroy the salmon of a district,
and how difficult it is to restore the breed.

At once the most primitive and most deadly
method of catching fish which inhabit rivers is by
the erection of built barriers and enclosures. The
Australian blacks make dams and pools of branches
and stones, and drive the fish into them before floods
subside. The North American Indian has for long

xvi INTRODUCTORY

years erected his barriers of logs to stop the ascend-
ing salmon, so that he may scoop out his little supply,
regardless of the consequences to the river stock.
The officials who supervise the salmon fisheries of
the American continent are still opening up those
Indian dams. The fishing weirs of England and
the cruive dykes of Scotland are the British pro-
totype of the aboriginal structures ; indeed, the
Gaelic word cruive seems to indicate that branches
were originally used. They are now stone weirs
with fishing boxes in the gaps usually the only
gaps left for the descent of the water and ascent
of the fish. The salmon yairs of the Solway Dee
are analogous structures in which the leaders or
walls are still made of wattle.

Cruive dykes were erected in most Scottish rivers
in early days, but where owners of such structures
held also interest in upper waters the dykes were
for the most part voluntarily abolished on account
of their extremely injurious influence on the breed-
ing stock. In 1424 they were abolished by statute
when set in fresh water " quhair the sea filles
and ebbis," and now only some six remain in Scottish
rivers, only three of which are fished.

In spite, however, of those ancient statutes pro
viding for the preservation of breeding fish, and of
the more modern regulations as to meshes of nets
and annual and weekly close times, as well as recom-
mendations for the reduction of legitimate netting,
there are those who regard such restrictions, such
interferences with the liberty of the subject, as
superfluous and undesirable. In evidence before

INTRODUCTORY xvii

Lord Elgin's Commission in 1901 I heard the asser-
tion made that salmon spawn in the sea as well as
in fresh waters, and I have met more than one
tacksman of salmon fishings who insisted that a
grilse was a different species of fish from a salmon.
These views die hard, especially when there seems
to be some commercial advantage in holding them.

Since the great development of rapid transit by
which fish can be sent to market, the use of ice,
and especially since the sharpening of competition
amongst salmon fishers and the consequent great
improvement in methods of fishing, a more exact
knowledge of the life of the salmon has become more
and more necessary. We not only require to know
when the fish breeds, and that it can only reproduce
its species in our rivers, but we want to know how
often the fish breeds in its lifetime or, perhaps one
should say, how seldom it breeds ; at what age it
begins to breed and at what age it ceases to breed.
We also want to know its habits of migration to and
from fresh water, and how much time it spends in
the sea ; whether fish migrate from one river to
another, and where salmon go in the sea, what the
salmon's food is and when it is in best condition.
More than this, for the better regulation of the
fisheries we want to know the conditions which
govern fish in ascending rivers and tributaries a
subject of value to netsmen and anglers alike.

These are definite points of practical importance
nowadays, for on a thorough knowledge of the life
history of the fish itself can sound regulative treat-
ment alone rest.

b

xviii INTRODUCTORY

In former days the legislator said in effect, " The
salmon breeds in our rivers, and after breeding is
not fit for human food: we must protect the fish
while it breeds, and protect the lieges against un-
wholesome feeding." The system of having District
Fishery Boards and the supervision of a central
authority has now sprung up, however, and our
legislation must necessarily take account of this and
of the duties which naturally fall to be dealt with
by one body or the other. In Scotland, our District
Boards are composed entirely of the proprietors who
own the rights of fishings. The Crown may be, and
frequently is, a local proprietor of such rights, but
none but proprietors have apparently any say in the
matter of local management. With the natural
trend of modern development it is perhaps fair to
assume that in the future District Boards will have
assigned to them wider discretionary powers : they
will be expected to know what is for the best interest
of the districts under their charge; while the central
authority will be expected to concern itself in the
general interests of the salmon fisheries as a whole
and in the particular questions which, arising locally
and of local importance, require settlement at the
hands of a neutral authority.

Already in many districts private associations
are springing up in order to deal more effectively
with such points, for instance, as the amount of
netting which safely may be allowed in particular
districts, a point which, if they had power, District
Fishery Boards might very well be expected to con-
trol. Each district has to be dealt with on its own

INTRODUCTORY xix

merits in a case of this kind. These associations are
usually, it is true, concerned very largely in the pre-
servation of good salmon angling. But it is to be
borne in mind that this very preservation of good
angling plays a most important part in maintaining
an adequate breeding stock of fish. The number of
fish captured by rod, although in some districts it is
very considerable, does not seriously reduce a river's
stock as indiscriminate netting does. The most
exhaustive inquiry into the welfare of our salmon
fisheries conducted by the Elgin Commission and
reported upon in 1902 resulted in, inter alia, a most
important recommendation in favour of reducing
netting in narrow waters where fish congregate after
leaving the sea. It was decided that there exists
in every river a point above which it is advisable
that no netting of any kind be carried on, and that
the effort should be to allow a proportion of every
run of fish entering a river to ascend to the unnetted
waters.

The increased value of salmon angling makes it
possible, moreover, to compensate most fully those
who may be required to cease netting, and without
any hesitation it may be said that the result of action
of this kind has already proved most satisfactory to
all concerned. The Aberdeenshire Dee is a con-
spicuous example. The annual value of the salmon
fishing of the district has risen from 7000 to
19,000, and the Aberdeen Harbour Commissioners,
a body possessed of valuable salmon fisheries on the
coast and in the mouth of the river, and who fish
for commercial purposes alone, voluntarily subscribe

xx INTRODUCTORY

to the funds of the Angling Improvement Associa-
tion, because of the benefit which they find has
resulted to their own fishings.

To the angler also who is sportsman in the best
sense of the word belongs much of the credit of
having noted the special features of the salmon's life
in fresh water. By his experience he accumulates
knowledge of particular runs of fish as the netsman
at the mouth no doubt may also do, but he, as he
quietly casts his fly or spins his bait, hour after
hour and month after month, has the opportunity of
coming into closer touch with the many interesting
habits of ascent and distribution : he has his line in
the water before the netsman has a chance, and he
sees the season out after the nets are off; he is able
to note the peculiarities of habit in spring, summer,
and autumn. It is sometimes said that the charm of
angling is its uncertainty, but this takes account only
of the fish -killing results. The true disciple of Izaak
is a contemplative man who has his mind occupied
by many piscatorial reflections as he plies his art : he
is quick to observe, yet many perplexities make him
slow to draw his deductions. It is very often true
of him that the more he has seen the more he realises
how little he knows. Conversely, those who have
not seen very much are usually able to talk a great
deal about what they do not understand. Hence the
not altogether unfounded belief that fishermen are
cranks.

No doubt the primary object of the angler is to
catch fish. He goes forth equipped for that purpose,
eager for its fulfilment. But if there were no other

INTRODUCTORY xxi

attractions than the hooking, playing, and landing
of the salmon, some of us who do not catch many
fish because we always fish on a bad day, when the
river is too low or too high, or rising or falling, or, to
put it shortly and with much more truth, those of us
who are bad fishermen would find small pleasure in
angling. " Wha's catchin' fush ? " the small boy
retorted when rebuked for catching fish on Sunday.
He was fishing, and perhaps enjoying himself in spite
of a guilty conscience, but he was not catching
fish. In salmon angling there is the exercise and
the pleasure which belongs to the skill required in
the proper performance of the exercise, two elements
which form perhaps the chief joy of athletics and
outdoor games. But in addition there is the en-
grossing study which belongs to all forms of hunting,
the study of the habits of the creature hunted, so
that skill already acquired may be turned to greater
advantage.

From this the step is a short one to a more intel-
ligent knowledge of the life history of the creature
apart from its mere sporting value. It is very
natural, therefore, that sportsmen should be natur-
alists, and should render in the case of the salmon
valuable help to the elucidation of the fish's life.
Then the freshness and the leisure of life by the
river-side gives the characteristic setting to the
whole. How varied are the phases : the precious
grey day with the favourable breeze, the bluster
and battle of the day of squalls, when the whole con-
centration of the fisher and his every hope of success
is in self-control and quiet determined effort ; or the

xxii INTRODUCTORY

peculiar joy of standing waist deep in a clear High-
land river, the water lapping one's bare elbows as
one casts, the nearness of the glitter on the surface,
the sense of being part of this moving, life-giving
force ! These are joys, apart altogether from the
success of the sport, which more than compensate
for the strokes of ill fortune, the dire disasters when
sun, moon, and stars seem to fall from their places,
as well as for the petty exasperations which on
certain days seem constantly to dog one's steps.

But the angler carries away with him when he
leaves the river the results of his observations, and
when he smokes his evening pipe those salmon
problems rise again in his mind, and he goes to the
river on the morrow with fresh suggestions and
explanations, fresh points to follow out by renewed
observation. Sir Joshua said that to get his fine
results he mixed his paints with brains. So in sport,
as in other things where perfected appliances mean
a good deal, it is after all "the man behind the
machine that counts." Yet what a long time it has
taken us to gain anything like a satisfactory know-
ledge of the salmon.

It is only seventy years since Shaw,* at Drum-
lanrig, proved that the parr is the young of the
salmon and not a small adult member of the salmon
family. All the early writers, up to Parnell (1838),
describe the parr under the name of salmo salmulus,
and the author named states at considerable length

* Shaw, "An Account of Experimental Observations on the
Development and Growth of Salmon Fry." Trans. Roy. Soc., Edin.,
vol. xiv. p. 547, 1840.

INTRODUCTORY xxiii

the differences between the parr and the young
salmon. Those of us who have grown up since
Shaw's time can scarcely understand how the parr
could have been regarded as a distinct species, far
less the extraordinary controversy which the dis-
covery was the means of starting. Yet those early
disputes gave the stimulus needed for investigation
over a wider area. Since the earliest attempts at
salmon marking, at artificial culture, and the rearing
and crossing of different salmonids, as well as by the
study of the different runs of fish in our rivers, and
by the results of netting salmon on our coasts, a
body of information has slowly been accumulating.
Disjointed and very imperfect as this information
has undoubtedly been, it has nevertheless served to
indicate the lines along which more systematic infor-
mation should be sought.

A signal endeavour to crystallise our views in
this particular was the publication of Mr. Willis
Bund's " Salmon Problems," a book which has un-
doubtedly done a great deal to advance the genuine
search after radical facts in the life of the salmon.
Since this book was penned the doings of the salmon
have been followed with greater precision, and the
fish has, as it were, been made to tell his own tale
to a greater extent. The reliable identification of fish
caught, set at liberty, and recaptured has been the
means, through the instrumentality of the Fishery
Board for Scotland and the Board of Agriculture
and Fisheries in Ireland, of providing us with an
amount of information as to migrations and increase
of weight hitherto unapproached. Investigations as

xxiv INTRODUCTORY

to the feeding habits and nourishment of the salmon,
as to the records of growth, migrations, and spawn-
ing to be read from the study of the scales, as well
as inquiries into allied and auxiliary subjects, have
greatly added to the data now ready to hand.

My endeavour in the present volume is to draw
together results which we have received during those
recent investigations, but which have been set forth
in isolation, not set forth at all, or, what is perhaps
much the same in result, printed between the blue
covers which so successfully conceal a Government's
information.

CHAPTER I

SMOLTS

Early ideas concerning smolts Rearing of smolts in
captivity Marine versus fresh water origin of the salmon
Hatching fry from smolt eggs British smolts descend
when two years old Marking of smolts Habits in migrating
to sea Autumn migrations Development of smolts in sea
Herr Dahl's capture of smolts in the sea off Norway

THE Stormontfield experiments on the Tay, although
they at first gave new vigour to those who dis-
believed Shaw, ultimately, after the long dispute
described by Russel in his book on the salmon, be-
came a vindication of the Drumlaririg results and
a source of much additional information on the
growth and migrations of parr and smolts. Many
anglers of renown in the present day are puzzled to
tell a parr from a young trout. Yet, curiously
enough, the dispute never dragged the trout into its
vortex. The parr was either a young salmon or
a distinct species. Professor Grassi of Rome slew
the leptocephalus* idea without question. People did

* Leptocephalus was long known as an interesting semi-transparent
little fish not uncommon in the Mediterranean. The investigations
of Professor Grassi into the deep-water fauna of the Straits of
Messina led to the discovery that this little fish was simply the
larval or post-larval stage of the common eel.

A

2 THE LIFE OF THE SALMON

not seem to mind so much when only an eel's
reputation was at stake ; to tamper with the
salmon's genealogy was by comparison like dis-
puting a title to the peerage. However, the results
of the Stormontfield investigations were finally given
to the world in two small books, one by William
Brown in 1862, the other, since republished, by
Robert Buist in 1866. Buist superintended the
operations carried on by Peter Marshall, and both
were familiar writers to the angling papers of the
time, the latter under the title of " Peter of the
Pools."

With regard to the descent of the parr and smolts
from the ponds, the results were regarded as show-
ing that 60 per cent, descended when two years old,
32 per cent, when three years old, and 8 per cent,
when only one year old. This was indeed regarded as
the proportions in which the young salmon migrated
to the sea, but the ponds are several miles above
Perth and many miles from the sea, and it is safer
to say simply that in this proportion they descended
from the ponds. I have even heard some curious
hints as to the way certain fish were made to
descend, and we know that the argument was as
keenly contested as any land reform bill, but
such hints probably do not matter much. An
attempt was also made to rear smolts in salt water,
and a pond on the coast was secured for this
purpose. Unfortunately some local poacher put an
effectual closure on the experiment by stealing the
fish when, as was reported, they had attained the
size of herrings. The poacher not only committed

SMOLTS 3

an offence as serious as the time-honoured crime of
sheep-stealing, but robbed the observers of informa-
tion which might have saved them from much
subsequent error.

Various batches of smolts before descending from
Stormontfield were marked by removing or mutilat-
ing in some particular way the adipose fin. A
limited number were also marked by the attachment
of small rings. About the same time (1864) the
Tweed Commissioners were conducting some valuable
experiments of a similar kind. None of the Tay sraolts
marked by the attachment of the rings were recovered,
but three at least of the Tweed smolts which had been
marked by the attachment of wire were recovered as
grilse in the Tweed after a year's absence, one in 1855
and two in 1856, the first weighing 3^ lb., the second
and third weighing each 6^ lb. In spite of these, how-
ever, it was contended, with the publication of many
particulars, that the Tay smolts marked by fin-cutting
were freely recaptured as grilse of three pounds and
upwards, after intervals varying from one and a half
to three months, and as a result it became very
generally believed, in Scotland at least, that the
grilse of our rivers caught in May, June, and July,
or later in the season, had only descended as smolts
in the spring of the same year, and were therefore
only 2 years old. Other and similar fin-cutting
experiments were made elsewhere in Scotland, with
apparently the same results ; and illustrations accom-
panied certain of the reports to show there could be
no mistake about the conclusion. Curiously enough,
in the last few years a similar series of observations

has been made in England, and the results deduced
from them have been to show that the grilse had
been a whole year away from fresh water. I refer
to the experiments conducted in the Tavy in
Devonshire by the direction of the Duke of Bedford.
Here the method of marking was the same, but the
deduction different. I am not aware what the
appearance of the mutilated fin was in those
Devonshire grilse, but in Ireland, where some adult
fish marked with a label on the dorsal fin were also
marked " by a wide notch cut far into the adipose
or dead fin,"* it was ascertained that when one of
those fish was recaptured " the fin had recovered its
usual form, the outline of the notch being just
visible as a faint scar." With smolts kept in con-
finement, after mutilating the fins it has also been
found that the mark healed up, so that there was
considerable ground for the belief held by not a few
that this method of marking without the attachment
of some foreign substance was not reliable.

A good deal of additional information as to the
growth and migration of smolts has been obtained
by the numerous hatching and rearing operations
which have been carried on all over the country. It
is undoubtedly the case that the artificial conditions
materially affect the issue, and that in many respects
artificially reared smolts behave differently from those
in the wild state. The steady and abundant food
supply increases the rate of growth, and apparently
exaggerates the variation which may exist. For

* Holt, "Report of Sea and Inland Fisheries of Ireland for 1901,"
part ii. p. 181

SMOLTS 5

instance, I have two smolts which were reared at
the Duke of Richmond and Gordon's ponds near
Fochabers at the mouth of the Spey, which are ot
precisely the same age, viz., twenty-one months. One
is fully eight inches in length, while the other is only
four inches. Under natural conditions experience
shows that it is well-nigh impossible to find a salmon
smolt of eight inches even at twenty-six months.
The smolt of the sea trout may occasionally be found
of this size, but not that of the salmon. The reason
for the greater size of the sea trout is simply that
this fish has a much more estuarial habit, and after
descent of the fresh water continues to feed and live
and apparently to move up and down in the tidal
waters of estuaries. Many common brown trout are
to be found under precisely similar conditions and
feeding on a purely marine diet ; in the estuary of
the Tay near the Tay Bridge I have found large
silvery common brown trout with freshly swallowed
herring as well as much digested herring in their
stomachs and intestines. Yet we must not too
hastily believe that we see in such instances as
these the cause of the seaward migration of the
salmon. In spite of the fact that the eggs of the
salmon can only live and develop in fresh water,
there are many who consider that the salmon and
indeed all salmonids are originally of marine origin.
Mr. Boulenger, whose word on such matters is im-
portant, points out, in his chapter on the salmon
family in the Fishing volume of the Country Life
Library, that " the overwhelming majority of the
members of the sub-order of which the salmonids

6 THE LIFE OF THE SALMON

form part inhabit permanently the sea, the clupeids,
or herring tribe, which are their nearest allies, being
certainly of marine origin, as proved by their abun-
dance in cretaceous seas, yet a few, like the shads,
ascending rivers to spawn without this ever having
been adduced as evidence in favour of a fresh water
origin of the family to which they belong."

The more we learn about the salmon, the more we
have to realise the very considerable amount of time
which it apparently spends in the sea. Sea trout
are chiefly estuarial in habit. Loch Leven trout
may very possibly be sea trout, or at any rate
migratory trout which amounts to much the same
thing shut off from the sea. It is easy by a little
judicious feeding to make them practically indis-
tinguishable from ordinary sea trout. Brown trout
taken from Dorsetshire to New Zealand quickly
acquired a migratory habit and became large silvery
fish, inhabiting the sea for the most part, and ascend-
ing rivers to spawn. The brown trout which I have
referred to at the mouth of the Tay were clearly
doing the same, while in localities such as Orkney
and Shetland and the Outer Hebrides we have the
established tidal variety which has been called S.
orcadensis ; and in the West of Ireland we have
S. estuarius, the so-called slob trout. I do not agree
that there is any specific distinction, any more than
I agree that a ferox is not a brown trout. In all
these examples we may say that to feed is in one
sense the impulse which causes the change of habit
and consequent modification ; and in the same way,
taking salmonids as we now find them, it is evident

SMOLTS 7

that artificial feeding can still further induce very-
marked differences.

In all rearing ponds in Scotland, however, the
smolt assumes the migratory dress at two years of
age, although it commonly attains to the size of
the natural or wild smolt before that period. This
is strong presumptive evidence that the smolt
naturally lives and feeds in our rivers for two years
after hatching, and that it then is ready to go to
the sea. Several attempts have been made to pre-
vent the descent of smolts, so as to produce land-
locked salmon. In most cases the attempts have
failed, owing chiefly to the disappearance or death of
the fish, as in the experiments of Messrs. Morgan and
Pell in South Wales and of Dr. Murie in the ponds
of the Zoological Gardens. At Howietoun greater
success ultimately attended experiments made in
this direction, for although many smolts jumped
from the ponds and perished, so that the ponds had
to be netted over to preserve the remainder, a fish
was ultimately reared which was fertile and which
spawned, and from which fry were reared.* The
resulting specimen, however the one fish which
spawned which was described as a grilse, was a
very poor apology for a grilse as we properly under-
stand this sprightly young salmon. The largest fish
which is figured is ill-shaped, spotted, and with the
parr marks still evident, and weighed only Ij Ib.
It may resemble in some measure the land-locked
salmon of the American lakes or of Sweden, but it is
a poor thing to call a grilse, and forms in my opinion

* Day, Trans. Linn. Soc., 2nd ser. " Zoology," vol. ii. part 15.

8 THE LIFE OF THE SALMON

a strong argument in favour of allowing all smolts to
get to the sea. It is much easier to accustom a young
parr or a Loch Leven trout to salt water than to
accustom a smolt to fresh water.

It has to be noted, however, that the eggs of the
fertile females of the Howietoun experiments were
not fertilised by milt from male salmon reared
under the same conditions, but by milt from Loch
Leven trout. This materially detracts from the value
of the result. Within recent years a limited number
of the smolts reared in the Duke of Richmond and
Gordon's ponds at the mouth of the Spey were found
to contain well-developed ova, and an experiment
was made as to the possibility of rearing brood from
smolt eggs. It has long been known that male parr
or smolts are occasionally sexually mature. So far
as I am aware, the occurrence of ripe female smolts
amongst artificially reared and hand-fed salmon had
not previously been noted. Aripe male smolt not being
procurable when the ripe females were first found,
fertilisation was satisfactorily accomplished by using
milt from an adult salmon. The fry hatched out and
grew quite normally, and, as I was able to note from
specimens kindly sent me by Mr. Kae, the superin-
tendent of the Duke's salmon fisheries, were at the
age of one year quite indistinguishable from the
salmon fry produced from the eggs of adult
females.

The following year, however, when ripe female
smolts were again discovered, care was taken to
obtain ripe male smolts for purposes of fertilisation
On this occasion, although at first impregnation

SMOLTS 9

seemed to have been successfully obtained, all the
eggs died before hatching.

In the experiments above related the abundance
of food was carefully seen to, and up to the smolt
stage, as has been said, such artificial feeding will
induce excessive growth. Yet in attempts to pro-
duce grilse in fresh water the best result is a stunted,
flabby-looking fish. But a few attempts have also
been made to produce grilse in confinement in salt
water ponds. We have already seen what became
of the Stormontfield fish. A fish from the Usk, kept
in the salt water tanks of the Brighton Aquarium,
is reported to have lived for five years, and to have
then attained to a weight of 8 Ib. Some similar
experiments have, I understand, been made at Ply-
mouth in the laboratory of the Marine Biological
Association. Other experiments have been made at
the mouth of the river Spey, in a pond specially
excavated for the purpose by direction of the Duke
of Richmond and Gordon. Those last-mentioned
experiments I have in a measure been able to follow-

The pond was constructed of cement work sunk
in the gravel of the sea beach, and was connected to
the sea by means of a syphon. In the second week
of September 1901 some twenty-five smolts nearly
two and a half years old were first introduced. In
the adjoining fresh water ponds they had been raised,
and having been retained beyond their natural time
of migration, had already assumed and then lost the
silvery dress. They were put straight into pure
sea water, where they ceased feeding at once, and
where in a few days four died. On a certain admix-

10 THE LIFE OF THE SALMON

ture of fresh water being allowed to enter the sea
pond the fish recommenced feeding, and by regulat-
ing the density of the water so as to obtain a more
gradual transference from fresh to salt, the fish
became thoroughly acclimatised. Owing, however,
to an unfortunate accumulation of disintegrated and
decaying seaweed drawn in from the outer beach
through the syphon, and which was suddenly stirred
up so that a large quantity of gas was liberated,
many of the fish suddenly died, and the remainder
had to be transferred to the fresh water ponds.

After a second attempt had been made under
healthier conditions, Mr. Muirhead, his Grace's com-
missioner, kindly sent me a specimen 13 inches long
(33 cm.). It was a beautiful silvery little fish, and
had been reared for two years in fresh water and
one year in the sea pond. It was figured along with
a specimen of a young fish from Galway river in the
Proceedings of the Royal Society of Edinburgh
(xxv. p. 395). I understand, however, that for
some reason or other the pond has again become
unsatisfactory ; but it would appear that, given
healthy conditions of sea life and suitable food, the
artificial rearing of grilse is quite a possibility.

These experiments under artificially induced con-
ditions do not, however, teach us a great deal about
the real life history of the salmon. The descent of
the smolt to the sea is prompted by a most powerful
instinct. Give fish in confinement as much and as
carefully selected food as they can eat, and the
silvery smolts are in no way induced to forego their
seaward migration. For the healthy growth and

SMOLTS 11

development of the species salmon as we now know it
the bracing qualities of the sea, with its rich feeding,
are absolutely necessary. I am aware that some are
inclined to insist on the importance of the amount
of feeding in our streams, and to go the length of
suggesting that the stock of a salmon river should
not be allowed to increase beyond the point where
the smolt food fails to go round. To this I would
reply that I would not limit the supply of adult fish
on the chance of the parr starving, first because parr
are evidently able to do a good deal of starving, and
secondly because no man can well estimate the
amount of parr food in a river. I should start to
kill off the trout first. Also it seems to me that
the amount of variation in the times and seasons
when parr can go into sea water is probably an
excellent provision against any danger of over-
crowding.

Herr Dahl, in Norway, considers that the parr
there leave the rivers when one year old. I confess I
am not satisfied that he is correct, and an examination
of the scales of Norwegian parr which I have had an
opportunity of seeing thanks to Mr. H. W. John-
ston supports this view ; yet we know that the
severe winter conditions of Norway modify very
materially the habits of the adult fish. The fry of
other salmonids seem to have different habits from
those observed in Britain. Mr. Rutter has investi-
gated the matter in the Sacramento river of the
Pacific coast of America, and finds that there the
fry begin to descend whenever they can swim, and
commonly reach brackish water in about three

12 THE LIFE OF THE SALMON

months' time. The adult fish are no doubt in great
numbers in those west coast American and British
Columbian rivers, yet the rivers themselves are of
vast size, and the amount of food for fry must be
practically unlimited. In any case we have no evi-
dence that smolts would starve in any of our rivers
if they did not descend when they do. No doubt
they get a greatly increased and varied amount of
food when they do go into the sea, but the time
when the migration takes place is the time when the
best feeding season is commencing, and it seems to
me necessary to take the natural instinct for a
temporary marine sojourn into account as well as
the need for increase of food.

The time at which spawning takes place naturally
influences the condition of the fry, and possibly,
within limits, the time at which the smolt enters the
sea. We have in Scotland at the present time, when
account is taken of the Border rivers, considerable
variation in the limits of the spawning season. So
far as reports show respecting the times at which
fish are noticed spawning in the various Scottish
districts which reports are published annually in
Part II. of the Fishery Board's Reports there is a
difference between the earliest and the latest mean
periods of fully two months and a half. If, however,
the Border rivers Tweed, Annan, and Nith are ex-
cluded and they have been subject to special and
peculiar conditions the difference in mean time
between Scottish districts is reduced to one month.
In other words, the height of the spawning season
in the earliest river, viz., November 7, is one month

SMOLTS 13

earlier than the height of the spawning season in
the latest river, viz., December 8. Of course, in all
districts one may find the spawning season extend-
ing over a long period if the very latest pair of
spawners are taken into account, and exceptionally
early or exceptionally late fish spawners caused, it
may be, in great measure by the presence or absence
of frost may account for a great deal of the diversity
seen in the sizes of parr at any one season, but the
chief periods of spawning have to be reckoned with
in estimating the period of the smolt's average stay
in fresh water.

The temperature of the water materially controls
the length of time between impregnation and hatch-
ing, and various experiments have been tried to
ascertain on the one hand how soon hatching may
be brought about and too early hatching is natur-
ally accompanied by a constitutional weakness of the
alevins and on the other hand how long hatching
may be retarded, so that eggs may be safely sent to
the Antipodes. Such points are of practical import-
ance to the fish culturist, but under natural condi-
tions in this country hatching may be expected to
take place in from ninety to 100 days, in tempera-
tures varying from 40 to 45 F. If the water
temperature is constantly kept on the verge of
freezing point (32 F.) hatching is retarded for about
148 days. If the eggs are frozen, or encased in ice
in a closed vessel, they die : oxygen must be taken
in and carbonic acid given off; the young salmon
even inside the egg must have the power to respire.
In like manner, if the eggs become coated with any

14 THE LIFE OF THE SALMON

deposit, be it natural sediment or polluting matter
of a more toxic sort, this transmission of gases
through the egg envelope is hindered or prevented,
with more or less serious results.

The salmon's redd or spawning bed is therefore
selected in clean" and not too fine gravel, where a
current of water brings a constant supply of the
necessary oxygen and prevents deposit. By succes-
sive quick movements of the salmon's tail the gravel
of the selected spot is displaced and with the aid
of the current it is surprising how large the stones
overturned in this way frequently are. A shallow
trough is thus formed, and in this the eggs are de-
posited. In nature all the eggs are not extruded
at once, because all the eggs in the ovaries do not
become ripe simultaneously. The eggs at the
posterior extremity of the ovary become ripe first,
and becoming free for extrusion are first deposited.
On extrusion they are at once fertilised by the
attendant male fish, and are then covered by gravel
moved as before. The female then leaves the redd.
As the remaining eggs ripen she returns and repeats
the process, till after the expiry of several days all
the eggs are shed.

It has repeatedly been noticed that the advent of
frosty weather induces the females to frequent the
spawning fords, and conversely a spell of mild
weather seems to prolong the period during which
spawning is carried on. Weather conditions there-
fore materially affect the length of time taken, but
I am inclined to estimate that on an average the
female salmon completes her reproductive functions

SMOLTS 15

in a week or ten days. In small streams the time
may be shorter.

When a salmon is stripped for hatchery purposes
as many eggs as possible are pressed from the
abdomen, and the fish is left in a collapsed and
wrinkled state quite foreign to the naturally
spawned-out kelt. At the hatcheries of the Pacific
coast rivers in America each female is first killed
and then slit open. A considerable proportion of
the eggs so taken are not naturally ready for
extrusion, and yet the percentage of loss through
unfertilisation or subsequent death is small. In
the natural process all extruded eggs are per-
fectly ripe, and it is extremely unlikely that the
percentage of unfertilised eggs is greater. Over
against this, however, we have to reckon with the
attendance of the hungry trout as the redd is being
covered, and the subsequent attacks of creatures
capable of penetrating amongst the gravel. With
the conditions which obtain in the great majority of
our Scottish rivers there is abundant evidence to
show that the natural spawning of the salmon pro-
duces a return in smolts sufficient to provide for a
full stock.

The egg, then, lodged in a dark crevice between
stones of the river bottom, develops till the time of
hatching arrives. The little fish, with its attached
yolk-sac, its food supply, then wriggles out, and for
a period of about fifty days continues to grow in
those dark recesses. Towards the end of this period,
as the yolk-sac has become much reduced, feeding
through the mouth is commenced, and efforts for

16 THE LIFE OF THE SALMON

exit to the upper water are begun. When the yolk-
sac has entirely disappeared the little fish is
practically an inch long, and is ready to roam over
the quiet shallows and to explore the sanctuaries
where such small fry congregate. Great numbers of
such tiny creatures may sometimes be discovered
in unexpected places. On one occasion I found an
undermined slope of a weir which I was examining
crowded with them ; on another occasion an isolated
pool round the pier of a viaduct was similarly
peopled.

At the commencement of their lives more than at
any other time the salmon move in shoals. They
are, of course, unable to withstand strong currents,
and are on emerging from the redds no doubt
washed down stream to the first quiet localities
where they can assemble, their instinct apparently
being to take up positions where they are not easily
reached by trout and other predatory fishes. In
such places it is to be feared, however, that certain
birds are most apt to find them, in spite of their
strong protective colouration. The black- headed
gull, which nests inland, and does not disperse
its young to the coast for several months, is a
pretty bird, which salmon fishers may freely regard
with pleasure in a lady's hat. He is an expert fisher
of shallow streams, and far outnumbers all other
birds which in my opinion must be regarded as
enemies of the salmon. He might well be excepted
from the protective list of many a County Council.

The inch-long salmon takes up its free-swimming
life towards the end of April, and has two years of

SMOLTS 17

river existence in front of it. Feeding is not
apparently carried on as steadily in winter as in
summer, if we may judge from the habits of parr
reared in ponds, and growth is not therefore
steady, but in a year the little fish is 85- to 4 inches
long. At two years the fish becomes a smolt, the
characteristics of which will be referred to presently.
All the rearing experiments show that in this
country the normal age for the silvery scales to
appear and for the instinct of seaward migration to
show itself is when the fish are about twenty- five
to twenty-six months old. Every one knows also
that the chief season for descent is the spring. The
old rhyme that the " first floods of May take all
the smolts away " is not far wrong. Climatic
conditions determine the amount of variation from
the normal, and we may state that mild springs, in
inducing an abundance of feeding early in the season,
seem to produce an early descent instead of a
late descent, as would be the case if the fish went
down solely for the purpose of procuring food. From
the end of April to the beginning of June is the
season of chief descent to the sea.

With regard to the late spawning times in the
Border rivers, to which reference has been made,
I would here venture to express the belief that we
have a condition brought about, not by any natural
cause, but by the action of man in regulating the
fisheries so as to allow an undue amount of netting
to continue too long each season. If the amount of
netting in confined waters was reduced and the re-
maining nets wisely regulated, the earlier running

B

18 THE LIFE OF THE SALMON

fish would be allowed to remain in better proportion,
and the spawning season would in time harmonise
with the normal. A long belt of netting in a
river renders the weekly close time inoperative :
all early fish are captured, and only late-running
fish remain to keep up the stock. The statement
is by no means speculative As will be shown at
greater length in a subsequent chapter, physio-
logical investigations of the most exhaustive kind
have shown that in widely different rivers geogra-
phically the ovaries of the salmon reach maturity
with remarkable seasonal uniformity.

Reference has been made to the great variation
in the rate of growth of artificially reared smolts.
This does not seem to exist in the natural state. In
a series of smolt-netting experiments recently
made in the waters of the Tay below Perth it has
been found that, while various sizes of parr are
present, the silvery or migratory smolt is wonder-
fully uniform in dimensions. Great numbers of
smolts were passed under observation, and batches
were measured at intervals. They were found
to be from 5 to 6 inches in length (measuring to
the centre of the forked tail), and to weigh from
1 to 2 ounces. The netting and marking of Tay
smolts was commenced in the spring of 1903.* The
part of the tidal water where the fish were found
in greatest numbers was about four miles below
Perth. The river above this point is affected by
the tide for several miles, but the fresh water is

* Twenty-third Report, Fishery Board for Scotland, Part ll.
Appendix III.

SMOLTS 19

merely dammed back by the inflow of the denser
water of the lower estuary. At Kinfauns, the part
of the river indicated, the smolts accumulated in
great numbers, so that with a small-meshed sweep
net it was possible without any difficulty to catch
them in hundreds at a time. We attempted to
follow the smolts in their further descent, and
proceeded down the river, drawing the net on one
side or other as we went. From the time we left the
neighbourhood of Kinfauns the smolts became fewer,
and when we had descended about two miles and a
half, and had reached a point a short distance below
the mouth of the river Earn, where sea- weed begins
to make its appearance upon the shore, smolts could
not be found at all. We proceeded down the estuary,
however, and, thanks to the courtesy of the Tay
Fisheries Company, who granted the use of their
steam yacht, completed a survey of all available
fishing places, both on the shores of the lower
estuary and on the shallow banks in mid stream
near the Tay Bridge, till eventually we reached
Budden Ness and the open sea, some twenty miles
below our starting point. Not another salmon smolt
did we catch, however, although sea trout smolts
were everywhere in evidence, as well as brown trout,
herrings, flounders, a young turbot, sand eels,
pipe fish, and various marine shore forms. In this
we repeated the experience of Herr Dahl in his
attempts to follow salmon smolts from the rivers of
Norway down the fjords.*

In June of the following year (1904) I was able

* K. D&hl, " CErret og Unglaks." Christiania, 1902.

20 THE LIFE OF THE SALMON

to make arrangements for a series of observations
by means of a boom-net used in fishing for sparlings
or smolts (osmerus). This net is worked from a
smack anchored in the channel ; the mouth when
extended by its booms and the pressure of the
current is 24' x 18'. Operations were commenced a
short distance below the mouth of the Earn, where
we had on the previous season lost trace of the
smolts, and were continued down the estuary at
intervals till a point was reached below the Tay
Bridge and opposite the city of Dundee, the essential
feature being the working of the net in the channel
or fairway. This fishing yielded most interesting
results. The net was set forty-six times, and salmon
smolts were now captured freely. Some 1085 young
salmon ids were taken, and the great majority of
these were salmon smolts, the others being sea
trout.

Two points were specially noticeable :

1. As the descent of the estuary was made, the
size of the salmon smolts did not increase.

2. All salmon smolts captured were taken when
the net was set so as to fish the ebbing tide. When
the net was set so as to fish the flood tide or incoming
current sea trout were alone taken, and many of
these were nearly twice the size of the salmon smolts.
Sea trout were also taken during ebb tide.

The natural inference from those observations
seems to be that, after congregating in great numbers
at the extreme upper limit of the brackish water,
the salmon smolts make a steady and comparatively
rapid descent to the open sea, without hanging about

SMOLTS 21

and growing in the estuary, as sea trout appear to
do, and that in making their descent they forsake
the shallows for the main current. In rivers which
flow directly into the sea with little or no natural
estuary, such as the Spey, Dee, or Helmsdale, the
transference of the smolts must be comparatively
rapid, but in such localities a considerable body of
fresh water no doubt spreads itself out delta-wise
over the denser sea water in such a manner that the
necessary conditions are supplied to the smolts. It
seems certain that just off or at the mouths of such
rivers the smolts are in great numbers in May and
June, and at times fall an easy prey to coal- fish and
other members of the cod family. The mouth of the
Spey has been occasionally netted when smolts were
seen to be descending, and great quantities of coal-
fish captured and destroyed or given away. The
stomachs of those fish were found to be full of
freshly swallowed smolts. It appears, however, that
the coal-fish do not always time their visit correctly,
since on more than one season of smolt descent a
very large draft net failed to capture any of them.

It has been said that in netting the upper tidal
waters of the Tay at Kinfauns parr of very varying
sizes are found. Some of those little fish are not
more than an inch and a half to two inches in length.
Without doubt many gravid adult fish entering the
river during the actual spawning season and it may
be stated here that ripe male fish come in from the
sea in full red spawning livery, as netting in close
time for experimental purposes fully shows that
these gravid fish make their redds and spawn almost

22 THE LIFE OF THE SALMON

within reach of sea water. The small fry found in
such localities may therefore be naturally regarded
as in part at least the offspring of salmon which
have spawned in the neighbourhood. It has to be
recollected, however, that two parts of sea water to
one of fresh kills salmon eggs. It seems probable,
however, that other young fry, following the instinct
to form into shoals which we know parr possess, have
joined themselves to companies of larger parr, or it
may be have been to some extent washed down by
floods. In any case these small fry and all stages of
parr are to be found in the upper tidal waters at all
seasons of the year.

In order to ascertain whether or not smolts
migrated in any numbers in the autumn, and at
other seasons than spring, netting at Kinfauns was
continued through an autumn and winter. This work
was taken up by the Tay Fisheries Company, with
the permission of the Tay District Fishery Board,
the catches being noted and the little fish returned
unharmed to the water. Murie, Day, Traherne, and
Willis Bund have all mentioned or described an
autumn migration as occurring in England or Ireland,
but evidence on this head was wanting from Scotland.
In the Tay the smolts were found in a somewhat less
brilliant dress during the months of September and
November 1905, but they were in small numbers.
On November 7, for instance, only twenty-nine were
taken after fourteen hauis with a small-meshed net.
The particulars and a detailed description of some
of the specimens are given in the twenty-fourth
Report of the Fishery Board for Scotland, Part II.,

SMOLTS 23

Appendix III. In December those silvery fish were
not found, nor was there evidence from the netting
during January and February that any migratory
smolts were present. On March 27, 1906, however,
fifty parr and smolts were taken, many of which were
about 5J- inches long (the extremes were 4*6 cm. to
14 cm.), and in appearance approximating to the
ordinary spring smolts. We have no positive
evidence, however, that those March smolts were
entering the sea, and it may be that they were
representatives of parr which, like the 8 per cent, of
the Stormontfield fish which were reported to have
descended after only one year in fresh water (the
ponds), have reached and have remained in those
upper tidal waters for some time, and which would
probably be amongst the earliest to leave for the sea
after a second year. This view is strengthened by
an examination of their scales made by Mr. H. "W.
Johnston, which showed that the fish were " rising
two years." Throughout this smolt netting I could
find no evidence that one-year-old fish left the river
for the sea or assumed an appearance suggestive of
their descent before they were two years old.

With the departure of the smolts for the sea we
may regard the first stage of the salmon's life as
completed. Up to this time the life has been in fresh
water, where observation on man's part is less diffi-
cult than in the sea. The earliest reappearance of
the fish in fresh water is as a grilse, and the study
of the intervening period of marine life has till
recently proved too difficult for most who have
attempted it.

24 THE LIFE OF THE SALMON

A certain number of stray specimens have from
time to time been captured, which have given us
glimpses of the salmon's development in this stage.
When trawling with a shrimp net for experimental
purposes in May 1901, two miles off Blackpool, i.e.,
about twenty miles from the mouth of the Kibble,
Mr. Archer was fortunate enough to catch a
young salmonid 7 inches in length, which Professor
Herdrnan and Dr. Noel Paton considered to be a
salmon smolt. Professor Herdman found the stomach
and intestines fully charged with nourishment. The
remains of two young sprats were in the stomach, and
the scales of other sprats or of young herring could be
distinguished in the con tents of the intestines, together
with the remains of small crustaceans (copepods),
which, however, may have been ingested from the
swallowed fish, since sprats and herring feed largely
on copepods. This smolt had apparently got clear
away from the mouth of any river, and had com-
menced to absorb the particular kind of nourishment
which we know the adult salmon fattens upon.
Herr Dahl reports * receiving from mackerel fishers
who had been working off the coast of Norway three
young salmon measuring respectively 17 '5, 39*0, and
43'0 cm. (6j, 15f, and 17 inches) ; and by fishing
with fixed engines of small mesh he himself secured at
the island of Garten, in the Trondhjem Fjord, a few
small salmon measuring from 45 to 70 cm. (I7f to
27^- inches). The same writer, who has made a most
exhaustive search in European museums for speci-
mens of young salmon, reports that in Bergen

* " CErret og Unglaks." Christiania, 1902.

SMOLTS 25

University are two examples, 23*5 and 28*0 cm.,
which Professor Collete informed him had been
found in Christiania fish market amongst some
young mackerel.

In 1904 there came into my hands a young sal-
monid, 14 inches long and 15 J ounces in weight,
which had been captured by Mr. Milne, salmon
tacksman, at the mouth of the Galway river, in
Ireland, and who was kind enough to send the
specimen for my inspection. I submitted the fish
to Mr. Boulenger, of the British Museum, who, along
with his colleague, Mr. Regan, agreed that it was a
specimen of salmo solar. I described and figured
the fish as such,* but pointed out the rather notice-
able depth of the caudal peduncle, and that in this
particular the fish resembled the condition commonly
found in sea trout. In the total length the small of
the tail (caudal peduncle) was contained llf times,
instead of from 15 to 15*9 times as in the smallest
grilse I have been able to collect in recent years. In
other respects the characteristics of the fish are
clearly those of salmon. In the way of hearsay
evidence one might mention a few similar occur-
rences of apparently stray specimens linking the
smolt and grilse. I may mention only one, which
the late Mr. Anderson, salmon tacksman in the Firth
of Forth and fishmonger in Edinburgh, communi-
cated to me, In his father's time, he informed me,
a fixed net used to be worked at Queensferry, and
while he was present, in June 1863, two small grilse
were caught each weighing f Ib. Instances like

* Proc. Boy. Soc., Edin., xxv. p. 395.

26 THE LIFE OF THE SALMON

these are no doubt exceptional, not only in that
they are rare, but also because such specimens taken
occasionally at or near the mouths of rivers at this
stage are apparently examples of young salmon which
have not followed the usual habits of their fellows.

To Herr Dahl we are indebted for the information
respecting the apparently normal habits of the sal-
mon at the stage we are now considering. As we
have already seen, certain specimens were obtained
by him in the sea at considerable distances from
rivers, and also taken by mackerel fishers. It has
been by following up this hint of the young salmon's
habit that he is now able to answer the often put
query, " Where do the smolts go after they leave our
rivers ? " Without doubt the satisfactory solving of
this problem is a most valuable piece of work, for
which the greatest possible credit is due to the
Norwegian investigator.

In the summer of 1905,* Herr Dahl renewed his
endeavours to obtain young salmon by offering re-
wards through the instrumentality of a Christiania
gentleman. He obtained twenty fish in this way,
captured between Sogne and Ulvasund. Eight of
these proved on examination to be young salmon,
the remainder being sea trout. The stomachs of
these young salmon contained remains of herring,
sand-eels, and sparlings. Permission to use a net of
small mesh was obtained, and an agreement was come
to by which a Mr. Waage of Brandasund should
employ it in mackerel fishing. Between January

* Knut Dahl. " Dyere Oplysninger om Unglaks og dens Ophold-
sfteder." Norsk Fiskeritidende, 12, 1906.

SMOLTS 27

19 and August 26 there were captured, between
Storholm and Brandasund, off the south coast of
Norway, forty-three salmon and 124 grilse. After
August 26 to the end of September, twenty-two sal-
mon and two grilse were obtained, giving a total of
191 salmon and grilse. Three of the grilse were
small, viz., 46 cm., 46 cm., and 45 cm. (18 inches).

In 1906 two fine-meshed mackerel nets were used
by Mr. Waage's men. One caught, between April 19
and August 26, twenty-three salmon and forty-seven
grilse, and five of the grilse were under 55cm. (21jin.),
viz., 37, 42, 42, 46, and 47 cm. After August 26,
twenty-seven salmon and four grilse were got, but
none of the latter were now under 55 cm. The other
net caught between July 3 and September 29 forty-
two salmon and grilse, but only one of the grilse was
under 55 cm.

In contradistinction to these results in the open
sea, Herr Dahl points to the results of his fishing in
the southern fjords in 1903, 1904, and 1905. He
made 930 attempts with a fine-meshed net, and in
this time only caught one specimen of salmo salar
under 55 cm.

His conclusion is that the young of the salmon,
after leaving the rivers and fjords, pass to the open
sea, and he declares that they occur constantly, and
can be freely caught if suitable nets are employed in
the natural habitats of young mackerel and herring.

We have already seen that our Tay sparling net
experiments showed that the salmon smolts seemed
to make a bolt for the open sea after leaving the
upper tidal water. We already know from the

28 THE LIFE OF THE SALMON

examination of sea caught salmon that herring and
sand-eels seem to be their favourite foods. Now
Herr Dahl has caught the growing smolts out at sea
as they follow the shoals of their marine "supporters."

Next we have to deal with the time spent by the
smolt in becoming a grilse before fresh water is
revisited.

From the seasons at which young fish in the
transition stage have been taken, as compared to
their size and weight, as also from the early Tweed
returns, which must be regarded as reliable although
so few in number, it seemed certain that the view
very commonly held in Scotland on this point was
erroneous, and that a full year in the sea elapses
between the time when the smolt leaves the river
and the time when as a grilse it returns.

To obtain direct and reliable evidence on this point
it was desirable to mark a large number of smolts
by the attachment of some foreign substance. The
presence of small spring fish in many of our rivers at
the same time raised a problem which seemed likely
to be solved at the same time and by the same means.
In 1903 I had experimented at Fochabers Rearing
Ponds as to a suitable mark for attachment. A
small silver disc was attached by a split silver pin to
the gill cover so that it lay flat upon the outside
and yet gave room for the necessary expansion.
The delicate bones of the smolt's operculum could
not, however, stand the strain of this mark, and in
a few weeks very many of the marked fish had
dropped both disc and pin by the rupture of the
gill cover.

SMOLTS 29

In 1904 in the lower Tay I tried a simple piece of
silver wire passed through the skin of the back close
to the adipose fin, but the wire chosen was rather
heavy. In the following year, however, Mr. Malloch,
Perth, tried a finer silver wire, inserting it through
the front of the dorsal fin and twisting it into a
simple loop, and this was found to answer very well.
From the commencement at Fochabers very great
advantage had been experienced in not handling the
smolts, but in using small vulcanite tubes for holding
the fish when inserting the marks. By using these
holders complete freedom from injury to the delicate
little fish through handling was secured. Under
Mr. Malloch's supervision, therefore, in 1905, the Tay
Salmon Fisheries Company undertook the marking,
and Mr. McNicol, who carried through the actual
operation, succeeded in inserting the wire loops in
6500 smolts. The admirable results which have
followed will be related in next chapter.

CHAPTER II

GKILSE

Return of marked smolts as grilse the following season
Habits of grilse in the sea Ascent of grilse from sea
Habit of many to remain in sea beyond grilse stage Small
spring fish The spawning of grilse Ascent of grilse when
in poor condition Argument in support of the marine
origin of the salmon Entrance of salmon to fresh water
not entirely explained by condition of satiety Reproduction
in fresh water Possible transference of small fish from one
river to another Movements of marked grilse along coast of
Scotland

THE smolts were all marked in the neighbourhood
of Kinfauns, where, as already explained, the little
fish can be most easily procured in spring. The
operations were carried on from the end of April till
early in June.

No recaptures of those wired smolts after attain-
ing the grilse condition were made during the
summer or autumn of the year of marking (1905),
although the tidal waters were netted day and night
as usual through the fishing season, and rod-fishing
continued till October 15.

On June 1, 1906, the first recapture was made,
the grilse having the unmistakable silver wire
attachment. Soon afterwards others turned up in
the nets. The first recapture was appropriately

c =

.

GRILSE

31

presented by Mr. Malloch to the Perth Museum, a
grilse of 2 Ib. 15 oz. The second specimen was
kindly sent to the writer, and was exhibited, with
the mark in situ, to the Royal Society of Edinburgh.*
It is a male grilse with genitalia in an undeveloped
condition ; length 24 inches, weight 4 Ib. 8 oz. The
interval of time between marking and recapture is
thirteen months. Between the date of the first re-
capture and the end of the fishing season, 1906, those
wired grilse occurred not infrequently in the Tay
nets. No recapture was reported from any other
locality. The weights varied from 2J- to 9 Ib. The
following details of date and weight respecting the
recaptures during June and July will sufficiently
show the results of this experiment as regards
grilse :

DATE. WEIGHT

Ib.

oz.

1.

June 1 . 2

15

2.

26

.

4

8

3.

28

4

12

4.

July 3

.

3

4

5.

4

.

5

8

6.

t

10

.

4

4

7.

j

14

.

5

12

8.

,

14

.

7

9.

16

6

8

10.

,

16

.

6

12

11.

t

16

.

4

8

12.

24

5

8

13.

.

26

7

14.

28

.

6

8

15.

28

.

7

2

16.

Aug. 4

5

" Note On the Stnolt to Grilse Stage of the Salmon," Proc. Roy.
Soc., Edin., xxvi.

32 THE LIFE OF THE SALMON

This is pretty good evidence that the first re-
appearance in fresh water of the fish which descended
to the sea as smolts about twenty-six months old is
not less than a year later when, as grilse, the fish
are three to three and a half years old. The
average weight of the Tay grilse mentioned is
5j lb., which gives an average increase, taking
fourteen months as the interval, of 6 oz. a month.
The increase in length may be taken as fully four-
fold. This rate of increase is surely more in harmony
with what we know of the growth of other creatures
having a duration of life similar to that of the
salmon than with the sensational increase demanded
by the view previously held, and stated as a proved
fact by several writers, that grilse come into our
rivers after an absence of only a month or two.

Reference has been made to the food of the Black-
pool smolt. In a little book, now out of print, on
the natural history of the salmon, written by a
salmon curer and tacksman of fishings at Inverness,
named Alexander Fraser, and published in 1830 at
the Inverness Courier office, I find this shrewd
observation : "I have found seven small fish in a
grilse of 3j- lb., and several, particularly herrings,
in the body of salmon." The writer then goes on to
speak of the sea as the feeding place of the fish :
"Salmon in lakes and rivers can, like other fish,
subsist for some time on little or no food. In the
river Ness they have been known to continue for
about eighteen months, the interval between their
entry and their departure for the sea, and during
this time their food must be very limited indeed.

GRILSE 33

It is in salt water that salmon obtain the chief part
of their food, and hence they fall off in condition in
proportion to the time of their abode in rivers."

We need not enter for the present on the vexed
question of the non-feeding of salmon in fresh water ;
suffice it to say that grilse, like salmon, as a rule
enter our rivers in a highly nourished state. They
follow shoals of herring to a considerable distance
from the mouth of their native river, and they show
remarkable cleverness, whether they swim in com-
pany or individually, in finding their way back
again. The fact that grilse are taken in bag- nets
set at considerable distances from the mouths of
rivers is sufficient to show that the wandering must
be fairly wide. Of course in the early stages of
their growth the fixed nets set on our coasts would
not retain them, since the mesh is a wide one
adapted specially for their safety. On the east
coast of Scotland the annual close time terminates
in most districts on February 10, and fixed nets are
again put in the water. Although during the first
of the season it is possible that the small growing
grilse would still pass through the mesh, we are, I
think, quite justified in believing that, were the fish
on the coast, grilse would be got in fair numbers
before May. The fact that grilse do not occur, or
practically do not appear, in the coast nets till May
leads one to infer that until about that time they are
not within the netted zone. As a rule in Scotland
the earliest grilse seem to appear on the coasts of
Banff and Aberdeen, where a few begin to appear
about the middle of April. The record for earliness

34 THE LIFE OF THE SALMON

is, I believe, March 6, 1905, when a 2j Ib. grilse
was caught at the mouth of the Dee. Willis Bund
mentions (" Salmon Problems," p. 87) that on rare
occasions grilse appear in the Severn in February.
Such occurrences are, however, altogether exceptional,
and fishermen do not look for a regular run of grilse
till May, while the greatest run is usually con-
siderably later. The special point of interest seems
to me to be that the grilse appear, as it were, all at
once on the coast, so that within, it may be, two
or three days such a market as Aberdeen may be
flooded with them, where only very few were before
to be found.

The year 1902 was a remarkably good one for
grilse, and having been favoured with a return of
the numbers of salmon, grilse, and sea trout caught
during the season by the Aberdeen Harbour Com-
missioners, I constructed charts showing graphically
the relative captures.* From the fixed net fishery
on the coast north of the mouth of the river Dee the
first records of grilse were on April 15, but com-
paratively few were taken between this and the end
of May and beginning of June. Signs of a great
run of grilse appear in the chart in the second week
of June, and on the 23rd of the month there is an
enormous catch. Good takes last well through July,
and at the end of July and again from August 10
to 13 there are two other marked rises in the curve
representing grilse. These last rises are also notice-
able because simultaneously occur rises in the salmon

* Twenty-first Annual Report, Fishery Board for Scotland;
Part II. Appendix II.

GRILSE &

curve, as if the particular influences which brought
grilse in shore to the netted area of the coast also
affected the salmon in the same way. A separate
chart of curves was given for the sweep-net fisheries
in the mouth of the river Dee. In this the first
grilse are shown as caught on April 18, or three days
later than on the coast, while the date of the greatest
catch is June 27, or four days later than the
maximum in the coast nets to the north of the river
mouth. The fixed nets, which are situated to the
south of the river mouth, do not share in the marked
influx of grilse. From these considerations it would
appear that the grilse approach the shore from the
north of the Dee, that they travel together in large
numbers, and, working along the coast in a southerly
direction towards the mouth of the river Dee, very
slowly and no doubt in an off and on manner (else
the first nets encountered would absorb the great
bulk of the run) enter fresh water three to four days
after having struck the coast " from the bosom of
the deep," as a fisherman picturesquely put it to me
on one occasion. Everything we know about the
salmon in this stage seems therefore to indicate that
after leaving the river as a smolt, which it does com-
pletely and with some rapidity, the young fish does
not as a rule again visit the shore, much less the
river, until a year has passed, when in company
with its fellows it swims into shallow water and
when some re-enter the rivers of their nativity.

In the Baltic around the island of Bornholm a
line-fishery exists for salmon. The bait used is
herring, and the lines are set so as not to go to the

36 THE LIFE OF THE SALMON

bottom. Isolated cases have occurred of salmon
being captured on cod lines off Scotland, and of steam
trawlers capturing salmon ten to forty miles from
land,* 1 and I have heard of the capture of a salmon
in the Pacific, some 400 miles off the American coast.
This was, of course, not salmo salar, but none the
less the occurrence is interesting, as showing that
salmonids do go far from land. Yet no marked
salmon has been caught after crossing the North Sea
or even St. George's Channel. One or two Nor-
wegian marked fish did, I understand, turn up in
England, but on investigation it was found that
they had come across the North Sea in a steamer,
just as one or two Scottish marked fish which were
discovered in a fish shop in Cornwall went from
Aberdeen by rail. In pursuit of their natural food,
however, both grilse and salmon clearly leave the
coast, and the fact above referred to that grilse
make their appearance in the coast nets somewhat
abruptly, and that a good " head of salmon " some-
times occurs simultaneously with this run of grilse,
leads to the inference that they have forsaken the
feeding grounds when they draw in to the land and
move along the coast. On the east and south coasts
of Skye and the east side of the island of Raasay the
bag nets catch such a large proportion of grilse that
the fishery may be spoken of as a grilse fishery.
On this account the nets are not put in the water
at the commencement of the season, but only
brought up to full strength before the main runs of

* Twenty-fourth Annual Report, Fishery Board for Scotland,
Part II. p. 76.

GRILSE 37

grilse may be expected. When the grilse appear on
the coast of Skye it is certain they are not on their
way to any rivers in Skye, because the Skye streams
are in summer so small that even a grilse would run
aground amongst the stones. When the autumn
rains have swollen those streams, fish ascend to
spawn. In the same way the Raasay grilse must
be on a passage elsewhere, since there are absolutely
no salmon streams in the island. The manager of
the fisheries says he has no idea whence the fish
come or whither they go.

But many of the grilse which travel along the
coast do not go into fresh water. When the chart
of curves illustrating the 1902 grilse fishing at
Aberdeen was constructed the net and coble fishing
was also recorded from the river mouth. The results
of this fishing in the river are far below the result
of the bag-net fishing on the coast to the north of
the river. It may be said in reply to this that the
coast nets met the first of the run and caught the
great majority of the fish. There is no means of
proving that they did not, but equally it may be
said there is no reason for believing that they did.
The fish were on the coast in great numbers.

The following statement shows the number of grilse
and salmon, and their weights, taken at the fishings of
the Aberdeen Harbour Commission from 1872 to 1897,
the great majority of the grilse being captured in
coast nets. The table is from the sixteenth Eeport
of the Fishery Board for Scotland :

38

THE LIFE OF THE SALMON

SALMON.

GRILSE.

Total

Total

SEASON

No.

Weight
inlb.

Wght. for
each
period of

Percent-
age.*

No.

Weight
inlb.

Wght. for
each
period of

Percent-
age.*

7 years.

7 years.

1872

3,981

59,303

5,923

27,338

1873

4,200

48,895

7,239

34,598

1874

8,100

88,398

6,187

27,219

1875

5,961

63,987

4,714

20,501

1876

2,873

33,788

9,184

39,297

1877

4.811

55,086

7,737

37,556

1878

8,918

116,308

465,765

3,350

14,446

200,955

1879

1,897

22,964

7,284

31,141

1880

4,178

43,587

5,410

23,720

1881

4,873

47,191

11,400

49,349

1882

6,530

62,717

7,415

32,413

1883

6,886

70,384

13,852

70,555

1884

5,424

64,866

6,129

27,817

1885

4,597

53,989

365,698

78-5

15,129

73,958

308,953

1537

1886

6,370

68,536

9,583

39,502

1887

5,052

55,535

14,847

60,644

1888

7,843

83,292

12,905

58,653

1889

5,623

60,573

9,336

38,379

1890

7,085

66,114

9,220

43,140

1891

9,477

107,566

11,775

58,318

1892

8,594

88,932

530,548

113-9

11,329

48,620

347,256

172-8

1893

6,795

71,185

5,068

18,131

1894

4,581

45,932

6,277

28,020

1895

7,002

75,603

14,938

76,836

1896

10,790

122,746

9,766

41,426

1897

4,834

50,218

4,060

16,948

* The weight for the first septennial period is taken at 1 00, and lor the later
periods as proportional parts of the first.

There are 229,957 grilse as against 157,275 salmon.

I have also figures for 1904, showing in like
manner the grilse and salmon caught on the coast of
Caithness in the Thurso district. In July of that
year there were 2032 grilse as compared with only
290 salmon. The Thurso is not conspicuous as a
grilse river.

The seine nets worked in the narrow waters of a
river are more deadly engines than the fixed nets on
the coast a statement which, I suppose, very many

GRILSE 39

will call in question. If the grilse had entered the
river Dee in any great numbers, the seine nets would
have found them. With regard further to the ques-
tion whether or not all the grilse come into our
rivers, I beg leave, with the recognised privilege of
a Scotsman, to answer this question by asking
another, viz., What is a small spring fish? If all
grilse enter our rivers, how is it possible that small
spring salmon of 4, 6, or 9 Ib. weight can be found
in our rivers just when grilse have finished spawning
and are descending again to the sea ? These small
spring fish cannot have spawned at an earlier stage.
If they had been up rivers on a previous occasion
they would have been very small grilse, and as kelts
would have been lighter than any kelts to be found.
But if I may anticipate a little here, I would say that
a study of their scales shows that they have not
spawned. Only two hypotheses are possible

(1) That they have previously ascended fresh
water, but have returned to the sea without
spawning.

(2) That they have never been into fresh water
since they left it as smolts, but have passed their
grilse stage in the sea.

With regard to the first, I would say that while
it is certainly true that some fish ascend rivers and
again drop back, there is no evidence to show that in
so dropping back any fish enter the sea for a renewed
period of feeding, but only through force of floods,
intense cold, injury or disease, or some more or less
temporary cause. Further, we have for several
winters carried on systematic netting operations for

40 THE LIFE OF THE SALMON

the purpose of marking fish, and at such times we
find grilse of both sexes continuing to ascend, and
invariably spawning or about to spawn. We have
never found any grilse which having entered the
river are to be regarded as non-spawners. The
second hypothesis, on the other hand, seems borne
out by all evidence of direct observation. The scale
of the small spring salmon is very like a grilse
scale in this, that a continuous sojourn in the sea is
stamped upon it, yet the fish seems to be the same
age as the grilse kelt. The scale of a 7 Ib. fish of
this class is figured in Plate III. of Mr. Johnston's
paper on " The Scales of Tay Salmon." * When refer-
ence is made to the tables showing the lines of growth
after smolt descent (pp. 70 and 71), we notice that
whereas in the second year grilse scales show in
May 7, in June 10-15, in July 14-21, and in August
17-25, i.e., an extreme range of 7-25 lines; the
small spring fish does not show less than 20, and in
February, March, and April has from 26 to 31 lines.
In other words, the small spring fish does not exhibit
a paucity of lines, which is readily explained in the
case of the grilse owing to the cessation of feeding
and the ascent of fresh water. This evidence is in
my opinion a strong corroboration of the view that
the small spring fish has not been in the river as a
grilse, but has passed the usual grilse stage in the sea.
In the spring of this year, however (1907), evidence
has been obtained which may be said to settle finally
the problem of the small spring fish. This valuable

* Twenty-third Annual Report, Fishery Board for Scotland,
Part II. Appendix II,

GRILSE 41

evidence is the recapture of small Tay " springers "
bearing in their dorsal fins the silver wire loops dis-
tinctive of the smolts marked in the Tay in 1905.
The first of these fish was taken on February 18, a
9 Ib. small salmon ; the second on the following day,
8 Ib. ; the third on February 21, a 9 Ib. fish. Others
followed through spring and summer until, in
August, after a stay in the sea of some twenty-
seven months, fish of 18 and 27 Ib. were captured.
The ascent of fresh water is for the first time. The
total number of those wired spring and summer fish
was considerably less than the number of the grilse.

Some of the light weights amongst these clean
fish are extremely like grilse, and this is to be ex-
pected, since they are only just passing into the more
advanced appearance, or the " gilling" appearance, as
it would be termed by Mr. Willis Bund. The grilse,
therefore, which we find spawning in our rivers are
only a proportion of the fish of their own age which
left the rivers as smolts (and which survived the
dangers of the deep). It may be this proportion
varies considerably in different years owing to the
greater or less inducement to remain in the sea, and
that this may be a factor in determining good and bad
grilse years, as much as the actual number of smolts
which descended the factor usually regarded.

The state of nourishment, the condition of grilse
entering rivers, certainly varies in different years.
I have already noted that 1902 was a good grilse
year ; it was good both in the matter of numbers
and the actual condition of the fish. Conversely,
1903 was a poor grilse year, the run not only being

42 THE LIFE OF THE SALMON

short, but the condition of the fish exceptionally
poor. I noticed some grilse in Dundee market
which were, without exception, the worst-conditioned
grilse I have seen. On inquiry I found they had
been sent to Dundee from elsewhere for very
shame, I imagined. In the same year some salmon
on the north and east coasts, more especially about
Caithness, were remarkably tasteless and dry in the
flesh, and at the same time I noticed that the local
herring tasted apparently of nothing so much as saw-
dust. Now the presence of those poorly nourished
grilse and salmon is noteworthy in two respects :
first, simply because fish were entering rivers in a
poorly nourished state (although some poor fish I had
evidence of were caught in sea nets) ; and second,
because the actual numbers of grilse were deficient.
As a rule grilse on entering the mouths of rivers are
found to be charged with nutriment, pyloric appen-
dages loaded with fat, and stomach empty. At the
mouth of the Tweed, Tosh and Grey * found 4 per
cent, out of 231 male grilse and only 1 per cent, out of
188 female grilse with signs of recent feeding. Hoek
at the mouth of the Rhine, examined 2000 fish, and
found food in only seven. The highly nourished con-
dition has been regarded as in itself sufficient reason
for the ascent of fresh water. Dr. Noel Paton puts
it thus : f " When on the rich marine feeding
grounds as great a store of nourishment as the body
can carry has been accumulated, the fish returns to

* Thirteenth Annual Report, Fishery Board for Scotland,
Part II. Note 2.

f " Investigations on the Life History of the Salmon," 1898,
Special Eeport, Fishery Board for Scotland,

GRILSE 43

its native element." The exhaustive and most im-
portant investigations of Dr. Noel Paton and his
colleagues demonstrated most clearly the manner,
inter alia, in which the great store of nourishment
is used up in the development of the genitalia, and
showed the corresponding loss in muscular tissue as
this development proceeds. With the condition of
satiety comes naturally the disposition to cease feed-
ing, and with the cessation of feeding the ascent of
fresh water commonly follows.

So far we seem to be on perfectly safe ground,
but I am inclined to think that the condition of
satiety has been somewhat forced in the argument
as to the result which it produces, viz., "the ascent of
fresh water. I admit that the normal condition of
fish on entering fresh water is that of high nutrition,
but let us inquire a little further into the argument
that this high state of nutrition causes the salmon
to leave the sea for the fresh water, or at any rate
that this is the primary and dominating factor in
the migration. Being in this highly fed up state,
fish are able to fast for long periods ; and be it noted
this is not peculiar to the salmon, but has been
ascertained to be true, for instance, in the allied
herring, whilst amongst higher animals the same
peculiarity is reported as being true of seals. Why
do they fast, and why does the salmon so fast only
in fresh water ? It has been said by way of explana-
tion that, having attained an overplus of nutriment
in its tissues, the salmon enters fresh water to rest,
and that this may be regarded as natural, since in
entering fresh water the salmon is in reality, as it

44 THE LIFE OF THE SALMON

were, returning home. It is essential for this argu-
ment that the salmon be regarded as a fresh water
fish which has gone to the sea to feed. This forms,
I admit, a plausible explanation of the presence of
early running fish in our rivers. That they do not
nourish themselves in fresh water has been proved.
Whether they can be said to feed in fresh water
must^not be discussed here. This point, about
which so much has been written, is really not very
important. It is in essence a dispute about terms.
Let us be thankful we can catch them with flies and
other baits.

But can the salmon be properly regarded as a fresh
water fish ? Dr. Noel Paton realises the importance
of this for his argument when he says, in dealing
with the factors determining migration, in the
general summary (p. 169) : "In considering the
question it must be remembered that the Salmonidae
are originally fresh water fish, and that the majority
of the family spend their whole life in fresh water."
To my mind it is not clear that the majority of the
family do so spend their lives, but further, it is not
sufficient to regard the British Salmonidse in this
matter ; we must take into consideration the group
of fishes, or at least the sub-order, to which the
Salmonidae belong, and judge the salmon as it
appears in relation to its near allies. If in this way
we take the seventeen genera of Boulenger's classifi-
cation, we find that the majority are more marine
than fresh water, including Baihylagus (Gthr), the
deep sea form recently added to the British fauna * ;

* Holt and Byrne, "Fisheries Sci. Invest, Ireland, 1905," II.
(1906,)

GRILSE 45

that one or two, such as Thymallus (the graylings),
are purely fresh water, and that the genera Salmo
and Coregonus include species which are both marine
and fresh water in habit. It may fairly be noted,
however, that the speciesof the two last named genera,
usually considered as inhabitants of fresh water alone,
can somewhat readily be accustomed to the sea that
indeed any salmonid may apparently be accustomed
to salt water whereas a pure fresh water fish, such
as a carp, at once succumbs to the effects of salt
water. Also we have Osmerus (the smelt or spar-
ling), which is classed along with four other genera,
three of which are marine, coming into our estuaries
to spawn. In the allied group of fishes, including
the herrings, we also see the herring itself, the sprat,
and the shads coming freely into estuaries, and the
shads spawning in fresh water.

If, further, we consider the salmonids, let us say
the British salmonids, which are commonly found
in fresh water, and note their characteristics in
relation to the common developmental characteristics
of the genus Salmo, we may observe that certain
features common to salmon parr, the so-called parr
marks and the complete dentition, very usually
survive throughout the life of the fresh water form.
If salmon smolts are retained in fresh water beyond
their natural time of descent to the sea, they assume
again the trout-like appearance of the juvenile.
We have already seen in the previous chapter that
the Howietoun specimens which spawned, and which
were figured by Day in the Transactions of the
Linnaean Society, retained the parr marks. The land-

46 THE LIFE OF THE SALMON

locked " salmon " of Lake Wenern, which seem in
reality to be trout, are reported to have not un-
commonly the same appearance. As the salmon
grows after visiting the sea, the teeth on the shaft
of the vomer bone disappear, and eventually the
teeth on the head of the bone go also. Trout retain
those teeth throughout life. In other words, the
trout commonly retains the juvenile characteristics
of the salmon, which develops fully only by visiting
the sea. This retention by fresh water trout of the
juvenile characteristics of the salmon does to some
extent suggest that the salmon has developed from
a trout-like origin, after acquiring the habit of
migrating to the sea. If this is so, Salmo trutta
and its varieties occupies an intermediate place
between the fresh water trout and the salmon. New
Zealand trout rapidly acquire the migratory habit.
But, on the other hand, and in view of the fact that
the majority of the salmon family are purely marine,
may we not have some cause for the view that our
fresh water trout itself represents a species in which
evolution has been arrested to a greater degree than
is seen in the case of the salmon. The salmon fall off in
condition after entering fresh water. It is easy to say
that they do so because they do not feed, but if fresh
water is their original element, why should they not
feed, and why should they fall off in condition ?

I firmly believe that many of us have got into
the way of regarding the salmon as a fresh water
fish simply because we observe him almost ex-
clusively in fresh water, where he spawns, and
further get into the way of regarding the silvery

GRILSE 47

scales of the fish as secondary and more or less
accidental results of his visiting the sea and feeding
there. We admit that he thrives surprisingly in
the sea for a fresh water fish, but we know he can
spawn only in fresh water, and ask how can he be
a marine fish if he cannot propagate his species in
the sea. Do we call the common eel a sea fish
because it spawns in the sea ? We call it the fresh
water eel. Is the shad a fresh water fish because
it spawns in rivers ? In my view, the prevailing
characteristics of the group of fishes to which the
salmon belongs are those of marine fish of plastic
nature, capable of much local variation both of
appearance and habit, many of which enter fresh
water freely. The fry exhibit a strong impulse to
descend the rivers of their birth and to enter the
sea; they do not develop into normal or healthy
adults if they are retained forcibly, nor under usual
conditions do they come to sexual maturity unless
they have visited the sea. In some localities where
great lakes occur a degenerate race becomes possible,
but these land-locked salmon cannot be regarded as
representatives of what the salmon used to be ;
they are not the fish through which evolution could
proceed. It is quite possible that Brachymystax,
the large salmonid of the Siberian rivers and Lake
Baikal, which viewed systematically comes between
Salmo and Coregonus, may be a migratory fish
which remains in fresh water for longer periods
than our salmon, and which in its great size and
full development differs in this respect from the so-
called land-locked salmon, but the life history of the

48 THE LIFE OF THE SALMON

fish seems still uncertain. It is significant, more-
over, that in these Siberian rivers the local species
of Coregonus, which with us are purely fresh water
fish (the vendace and powan), are fully migratory,
as reported by Giinther. It is quite possible to
imagine that in early times salmon were so numerous
in our rivers that a suitable supply of food was
impossible, and that therefore the habit of migrating
to the sea developed, but this is a mere supposition.
As a matter of natural history it is, on the other
hand, I think, more in harmony with what is known
of the habits of the other members of the group to
suppose that the habit which has developed is that
of migrating to fresh water primarily for the purpose
of spawning, and that in large continental rivers
and in rivers of land now disconnected from conti-
nents the habit of remaining for long periods, and
by some species of living more or less permanently,
has been acquired.

If, however, as I have shown, it happens that in
some seasons in our islands salmon ascend our
rivers in comparatively poor condition, a serious blow
is struck at the view that the primary cause of
ascent from the sea is the state of satiety. Along
with the procuring of food, the other great instinct
is that of reproducing the species, and a fish not
up to the normal standard of high nutrition may
still be impelled by this instinct, even if the repro-
ductive elements are, it may be, unable to produce
eight or nine hundred eggs per pound weight of fish.

I have dealt with this point of the marine as
opposed to the fresh water origin of Salmonidae,

GRILSE 49

because if the view in favour of the marine origin is
correct, our conception of the migratory habit must
be influenced accordingly. We must not regard the
descent to the sea, although it is the first migration
in the life of the individual fish, as the real direction
of migration, but rather that the parr is in fresh
water because the parent migrated from the sea to
spawn, and that the grilse and the adult fish alike
migrate primarily for the same reason, although the
condition of satiety plays an important part in the
case of early fish. When in a later chapter we deal
more particularly with the migrations of early and
late running fish the application of this conception
of the salmon's origin will again be taken up.

More inquiry is still needed as to whether or not
grilse and small salmon enter certain rivers, and in
later life as large salmon enter other rivers. In Alaska
it is reported, for instance, that small fish of very
uniform size are found in some of the smaller rivers,
while in large rivers of that coast, such as the Eraser
in British Columbia and the Columbia river in Cali-
fornia, only a much larger fish, but again of uniform
size, is seen. In a report for the year 1905 * Mr.
Archer suggests that further inquiry may show that
this obtains more often than is supposed, and in-
stances a case of "a fish marked in the Figgen
river in Norway a river in which the fish are
said seldom to attain to a weight of more than 1 1 or
12 lb., and are usually considerably smaller. This
fish weighed 5 lb. when first taken and marked,
and was recaught rather more than two and a

* Board of Agriculture and Fisheries, Annual Report of Pro-
ceedings under the Salmon and Fresh Water Fisheries Acts, p. xviii.

D

50 THE LIFE OF THE SALMON

half years later off an island on the coast more
than 90 miles distant from the river where it was
marked. It then weighed 18 Ib. The unusual

O

size it had attained, coupled with the fact that it was
taken much nearer to rivers frequented by large fish
than the Figgen river, is at least suggestive of
the idea that the explanation of some rivers holding
fish of a much smaller average size than others may
be due to the fish seeking new rivers after they have
attained a certain size."

In a chart showing migrations of salmon along the
coast of Scotland * are five lines indicating move-
ments of marked fish from the river Deveron,
which debouches on the southern shores of the
Moray Firth, to the east coast nets near Aberdeen.
Three of the five fish referred to were grilse, and in
each case this rather marked migration had been
made in four months or less. We have of course no
proof that those fish would not have travelled back
again, but I certainly consider such a return unlikely.

The great majority of marked fish, grilse as well
as salmon, are found again in the rivers in which
they were marked, but such cases are suggestive of
the possibility referred to above. Yet after the con-
siderable amount of salmon marking which has now
been carried out in Scotland, it is natural to sup-
pose that if such a transference occurred frequently
in our rivers, more direct evidence of it would by
this time have been obtained.

I have made reference to the large proportion of

* Twenty-fourth Annual Report, Fishery Board for Scotland,
Part II. p. 69.

GRILSE 51

grilse as compared with salmon which are caught in
many localities on our coasts, and to the fact that
grilse appear to affect some localities more than
others. Major Traherne tells us* that in the
Hampshire Avon and the Ogmore in Glamorgan-
shire grilse are absent, although adult salmon are to
be found. I believe, as a matter of fact, this state-
ment is open to question. It seems undoubtedly
true that grilse, although abundant on the coast, may
not enter certain rivers as salmon do. That, in
other words, the proportion of grilse to salmon on
the coast is a very different thing from the proportion
of grilse to salmon in any particular river. When
we come to deal with rivers, however, it has to be
remembered that we introduce not infrequently the
complications due to the action of man in, it may be,
overfishing or in polluting the fresh water, and that
changes may be going on which prevent our regard-
ing the results found as normal. In the Tweed, for
instance, the early reports show, according to Willis
Bund, that from 1808 to 1853 there were never less
than three grilse to one salmon, and that from 1853
to 1876 there were two grilse to one salmon. In later
years the grilse have continued to diminish. In the
Severn they have apparently become rather con-
spicuous by their absence.

Hoek at the mouth of the Rhine found that
grilse formed only 24 '1 per cent, of the fish ascend-
ing that river. Observations such as these are
necessarily made only during the fishing season,
while, as I have already said, winter netting in
" The Habits of the Salmon," p. 88.

52 THE LIFE OF THE SALMON

certain localities shows that grilse continue to run
with the latest of the spawning salmon. In the Tay
in 1902 and 1903 we carried on winter netting at
Almondmouth, a short distance above Perth, where
a very fair sample is to be obtained of the stock of
fish in the lower river. In November we netted for
three days in each year, and found a pretty even
balance :

Grilse. Salmon.

November, 1902 .107 . . 64
November, 1903 . 100 132

207 196

The fish enumerated are unspawned, spawning, and
spent. We caught in addition several clean run
salmon. At a higher part of the river, less suitable
for the finding of spawning fish, we had only eleven
grilse (all unspawned) and fifty-one unspawned and
spent salmon. I am unable to give figures worthy
of mention respecting the stock of fish in the upper
Tay, but in all probability the salmon considerably
outnumber the grilse, on account of the large number
of spring fish which ascend this well-stocked river.
Of no river of any size can it be said that it practi-
cally holds nothing but grilse, nor are the grilse in
any river with which I am acquainted in this country
in the proportion of seven grilse to one salmon, as
was the case on the Caithness coast in 1904. Willis
Bund refers to the small proportion of grilse in our
rivers, and adds that " the number of grilse, so far
as one can see, that spawn are below the proper
proportion." The same writer shows (" Salmon
Problems ") that it is impossible to show any clear

GRILSE 53

connection between good smolt years and good
grilse years. The explanation of this I have already
tried to suggest. Since all grilse do not enter fresh
water as such, one would not expect to find any
clear connection between good smolt years and good
grilse years ; the state of nutrition and development
of grilse in the sea being more likely to influence
the run of grilse to fresh water than the great or
small number of smolts which descended. The
summer grilse usually make a quick ascent, and
seem to show a preference for tributaries with rather
rapid water. The late running grilse, like the late
salmon, distribute themselves over the lower reaches
of rivers. After spawning it is remarkable how
quickly the great majority of them leave the river,
this being noticeable in grilse of both sexes, and in
large as well as small rivers. In our winter netting
in the Tay, grilse were found ripe amongst the last
of the late running salmon, and amongst descending
kelts they were amongst the first. During marking
operations it is common to retake salmon kelts which
have been marked and which are still hanging about
the same water, but we have hardly ever taken a
recently marked grilse kelt. Some grilse, however,
which have ascended a considerable distance appear
occasionally to remain longer than the average.

From the observations of Grey and Tosh * re-
specting Tweed fish, as well as from the observations
of Hoek and Meischer in the Rhine, it appears that
in July and August the weight of the genitalia of

* Thirteenth and Fourteenth Reports, Fishery Board for
Scotland, Part II,

54 THE LIFE OF THE SALMON

female grilse bears a much smaller proportion to the
weight of the body of the fish than is the case in
female salmon. In this respect grilse differ from
salmon, both large and small, where the proportion is
practically constant at 3*02. In grilse it is only
1*95 during the months named. In terms of 100
this percentage in grilse and salmon respectively is
as 64 to 100. As the season advances, the degree
of development of the genitalia steadily increases,
but the proportional development is later in the
grilse than in the salmon, or commences later,
although both grilse and salmon become fully ripe
at about the same time. These particulars were
clearly deduced by Archer * from the observations
referred to.

After leaving the river, it is clear that the grilse
can and does travel to fairly distant points on the
coast at considerable speed. A Deveron grilse kelt
marked at the mouth of the river (Duff House) was
found in a coast net 50 miles distant (Port Errol)
in 88 days. A similar recapture was made in 109
days. Another Deveron grilse which had travelled
70 miles to Cove, south of Aberdeen, in 122 days
was found on recapture to have gill maggots
(Lerneopoda salmonea, Linn.) still attached. The
fish was sent to me in Edinburgh with the mark
in situ, and I had the maggots carefully diagnosed
by Mr. Scott, a specialist in the lesser Crustacea
and a naturalist on the staff of the Fishery Board.
It is as yet uncertain how long those fresh water
parasites can remain attached to the gills in sea

* Fourteenth Annual Report Fishery Board for Scotland,
Part II. p. 9 et seq.

GRILSE 55

water, but presumably the fish's sojourn in the sea
had been comparatively short. A more striking
instance is to be found in a Brora marked grilse
the river Brora enters the sea on the east coast of
Sutherland which was recaptured in the Pentland
Firth, exactly 100 miles away by the coast line,
after an interval of only three and a half months
(109 days). There is some reason to suppose that
this fish had forsaken the river Brora, since a
Helmsdale fish marked within a few miles of the
other has also been taken in the same locality, but
instead of in a bag-net near the mouth of the river
Halladale, as was the case with the Brora fish, it
was taken by rod up the river Halladale. We
must not, however, overlook the fact that such
wanderers are in all probability exceptions to the
general rule, but, as has been said, there seems to
be a possibility that fish leaving small rivers, or
rivers only frequented by small fish, are more likely to
wander elsewhere than fish belonging to large rivers
such as the Tay.

The information as to hitherto unexplained points
in the life of the salmon, obtained by the sys-
tematic marking of recent years, must now, how-
ever, be followed out in a separate chapter.

CHAPTEE III

EESULTS OF SALMON MARKING

Early marking operations Modern method of marking
Divided migration Scottish and Irish records showing short
and long periods in the sea Increase of weight in kelts
recaptured as clean fish Significance of divided migration in
regulating fisheries Salmon returning to their own river
Recaptures in other rivers Recaptures on the coast
Direction of movements on coast Kelts recaptured as kelts
Spring fish marked and recaptured

THE immortal Izaak tells us in his " Compleat
Angler" (1653) of observations made on the homing
instinct of the salmon by tying ribbons or tapes to
the tails of young fish, and how the fish were found
to return from the sea to the same part of the same
river. More recent observers also sought to identify
fish so as to study their movements, but the methods
employed were not, as a rule, so harmless as those
of the lover of the gentle art. Mackenzie of
Ardross in 1823 attached brass wires to the tails
of salmon and grilse kelts. The late Duke of
Atholl used copper wire and a copper label the size
of a halfpenny in marking Tay fish, and between
1850 and 1863 obtained interesting recaptures,
although the copper, pressing tightly upon the

RESULTS OF SALMON MARKING 57

growing fish, is reported to have caused considerable
laceration.

From those returns we learned something of the
great increase of weight attained in salmon after
the sea has been visited, and of the length of time
certain individual fish had been absent from the
river ; but the investigations were not sufficiently
sustained to enable any one to understand the
variations which occur in any district or in different
districts and to make a sound deduction as to the
general habit and life history of the salmon in its
migrations, feeding, and reproduction. Still less
could proper deductions be made from other obser-
vations made by Young of Invershin, Fraser,
Buist, and others, whose method of marking was
almost invariably that of cutting the adipose fin of
the fish.

The one outstanding series of observations in those
earlier days was carried through by the Tweed
Commissioners, who from 1851 to 1864, and again
from 1870 to 1873, conducted a valuable set of
operations to elucidate the life history of the salmon
and bull-trout of the Tweed investigations which
were scientifically conducted as well as sustained with
regularity. These investigations have already been
referred to as yielding us the three early examples
of marked (wired) smolts recaptured as grilse.

The more recent marking operations carried on
by the Fishery Board for Scotland were commenced
by Mr. Archer during his term of office as Inspector
of Salmon Fisheries. They were instituted on the
lines of experiments already made by him in Norway

58 THE LIFE OF THE SALMON

for the purpose of discovering primarily the range
of influence of the Sands river. The treatment of
the fisheries of this his own river had resulted in a
very distinct improvement,* and the marking of
salmon kelts was commenced, so that the movements
of Sands fish could be followed.

The mark now used in Scotland is only slightly
modified from the pattern originally used by Mr.
Archer. It consists of a small silver plate, which
bears upon it a distinguishing number, and a silver
wire soldered to the long axis of the back of the
plate, and also passed through the plate so that the
two ends of the short wire project at right angles
to the plate. The mark is commonly attached to
the fish by pushing the two wire points through the
base of the dorsal fin, so that the surface of the
plate bearing the number rests upon one side of the
fin and the points of wire project at the other.
The points are then twisted together without
pressure upon the fin rays, if possible and bent flat
to the side of the fin. At time of marking a note is
taken of the weight, length, sex, condition, date,
and place ; and at recapture the same particulars
are, if possible, secured, and the records com-
pared.

The same system is now also adopted in Ireland,
where, however, a double plate is used, so that, if
necessary, the distinguishing number may be read
without the removal of the mark from the fish.
Both in Ireland and England the silver marks are

* Eleventh and Twelfth Annual Reports, Fishery Board for
Scotland, Part II,

PLATE II. Measuring- length of salmon on marking- board and piercing- dorsal
tin for insertion of mark.

PLATE III. Weighing- fish suspended by worsted tail band after insertion of
mark.

KESULTS OF SALMON MARKING . 59

now oxidised, so as to render them less conspicuous
to other fish.

Four reports on the Scottish results have now
been published (20th, 22nd, 24th, and 25th Ee-
ports, Part II.), and the records of over three
hundred recaptures discussed. In Ireland two
reports have appeared (Report for 1901, Part II.,
and Fisheries, Ireland, Sci. Invest., 1904, VII.
[1906]), and in England one short report has been
published (Agric. and Fisheries for 1905, Salmon
Acts, &c.). We have now, therefore, got a con-
siderable number of records with which we can
deal respecting the movements and increase of
weight and length of the salmon of our country,
and when we combine the results from the separate
countries we are in a position to view the salmon's
life with some breadth of horizon.

The most outstanding feature is the demonstration
of what has been called the divided migration of
the salmon the habit of the fish to remain short
periods and long periods in the sea. A shrewd fore-
cast of this dual habit as seen amongst Tay fish was
made by Mr. John Dickson, for many years agent
for the Tay Fishery Proprietors, and who wrote on
divided migration of the salmon as early as 1860
in the Perthshire Courier, and who has ever since
continued to assert that this habit when properly
understood will be found to completely explain the
presence of early and of late fish in Scottish rivers.
It has always been sufficiently clear to the observant
that all salmon are not annual spawners. The mere
fact that when fish are spawning, and when spent

60 THE LIFE OF THE SALMON

kelts are returning to the sea, we have clean, well-
fed fish with quite undeveloped reproductive organs
entering and ascending our rivers, is sufficient to
show that there are, at all events, a number of
salmon each year which disregard the reproduction
of their species. Such fish have sometimes been
described as barren, but this is simply because at
the time of capture they show no very evident sign
of sex or of the active function of their genitalia.
This missing of a spawning season has led other
observers to suppose that the salmon is a biennial
breeder.* The Scottish and Irish salmon marking
results show that the salmon is both an annual and
a biennial spawner, and from the study of the scales
there seems evidence for the belief that some fish
may spawn less frequently than even every other
year.

The salmon lives and grows in the sea, but has to
enter fresh water to spawn. It frequently enters
fresh water long before the spawning season. Fish
doing so in spring are fish which have spent what is
called the long period in the sea ; they have missed
a spawning season. Summer fish may also be fish of
long absence from fresh water, but very many of
them may be regarded as annual spawners, or at any
rate as fish which are reproducing their species on
two consecutive years. The understanding of this
feature this divided migration in the life of the
salmon is in great measure the key to " salmon pro-
blems." Both this spawning habit and its results,

* Atkins, " The Biennial Spawning of the Salmon." Trans.
Amer, Fisheries Soc., 1885.

RESULTS OF SALMON MARKING 61

and the growth of the fish with its many surprises,
must be viewed in strict relation to this divided or
dual habit of marine sojourn. It would be out of
place to compare at a given time a fish which had
returned to fresh water after a short sojourn in the
sea with a fish which had been feeding and growing
for twelve or eighteen months. When a classification
according to short and long periods is adopted, there
is still much that is difficult to understand and great
variation in results, but the life of the salmon seems
to be cleared of that obscurity which has for so long
defied the penetration of many and varied search-
lights. Moreover, it is also necessary to give atten-
tion to the localities from which fish come, since what
is considered a large fish in one place may be a com-
paratively small one in another, so that fish of similar
weight are not necessarily of similar class or age.
Further, it seems necessary to have regard to the
particular years in which records are acquired, and to
compare different periods, when treating the subject
broadly. Classification by weight, length, or other
particular is only useful in questions of migration
when fish of one locality are being dealt with. I was
never so much impressed with the necessity for com-
paring only similar periods till a combined analysis
of Scottish and Irish results was attempted. It has
been my intention in writing these chapters to avoid
tables of statistics, because of their rather formidable
and uninteresting appearance, but in the present
instance I must ask the indulgence of the reader in
giving the data in as condensed a form as possible
from which deductions are made.

62 THE LIFE OF THE SALMON

In dealing with the question of the short and the
long periods of sojourn in the sea, only one class of
particulars are reliable, the returns of fish marked
when kelts after spawning and recaptured as clean
fish in fresh waters or the mouths of rivers after the
termination of their sojourn in the sea. It happens
that in a combination of the Scottish and Irish
returns there are 117 recaptures of this class which
can be dealt with. In the following tables these
are separated into groups according to the years of
marking and recapture ; the records of fish marked
in certain rivers are placed together, and all fish of
short period are separated from fish of long period.
The details extracted are the numbers of the marks,
the intervals of time given in months, the increase
in weight, and the weights of the kelt marked.
Scottish and Irish results are given in parallel double
columns.

RESULTS OF SALMON MARKING

63

CO C5
O5 OS
05 OJ

!*

tf
LD

H

w

a

02
i I
H
H
O
O
02

oa

mill

W B K

X! OO <*< O> OO OO O

H

CIS i-H C^ t- t> O

05 <N (N

,-!-.-.

'fl'^OO

e< o >o

KW

64

THE LIFE OF THE SALMON

ft

SW

M

1

aj

G

PH

^

o

fc

3

|

02

i

M

B

1

H

C

rH

M

H

CO

HH

2

CO

o

o

03

3

1

(M

2

PH

1

O

S

ft

*

oa

CO

I

HHHHHHH

-

O

o

^

t*f

1

I ! 1

"" 1 1 I 1 II f 1 T-H

o
i

o

2

"3 u

i i r-5 i 1 1 i

cc

M

3>

no

1

So"S

-<JH OO OO M5 -^ O F 1

1

p

s M

H

H

o-gg

tf CO ~ ^H CC i I Oi
^ O O ^ O * ' C*l
O i i -H CM CO CO CO

PS

^

OO OO OO OO GO CC CO

B

OD

M

y

B

E

d o

t^ O O

C3o3c3c3c3cd O>S

00

E-t
H
O

CO

M

HHHHHH > >
OQ

O
02

|

o

*; -^

.-*M H. HN

3

**

c **

,gTt<OOCODOO ^lO3

H

CM

M ^

H

O

03

ss

.

i

*o> bo

X>5OOrHCC^?D -^O

o

08

M^

1 < i-H r-i

09

I

"^-^

HN

13

a M

otoi>osoo eoio

a
o
EH

64<

t^of^So^ CCN

!zi a

CO t** OO Oi O OO O) Od

RESULTS OF SALMON MARKING

65

a

m >>

-
60H

^

rr rt rt rt rr .. ^ .
fl3flfl><I)4 )aja;a3a '' 1>a) 'r2

aaadacaTLi=icoaaaW:g

66

THE LIFE OF THE SALMON

S

a
o

01

S

i

to

PH

*

O

5

i

s

"S

s,

5

V

_g

m

p*

-^

a

K
to

o

H

H

W

M

ec

a
o

o

03

3

T 1

co

-

PH

H

03

1

M
o

2

"

H

S

G

oo

h

IS

a

o

CM

BSSSSSSSfHHHHH

ec

3

^

PH ^^ td IT] fr"{ [T! [T]

P'

1

H

CO

g-sl

^ C J>. 1C t>-^5COI>"t'* < ^iOO50OOO

-

"3

PH

o

C5

i4

aS'a^wtJTS'^t.coa^w^

"a

o

w i i i i i (

3

TD

9

^ ^

a

*
M

1-sS

r^ooiooo>-woooo>t-^wo

"3

|

p

^ s

o

EH

o
o

i

^o|

i t i-Hi t^^^Hi li (OSOOOSOiOi

m

5?

b b -fa

H
H
O

I

a

fcfcfc HHH

CO

*.

o

O

^

CQ

O

O

.

.H ** HN

pi

^T*

c *o tc

, t^ i- 1-- T-H tO OO

^2

'-

PH

CD

o

M ^

u,

5

S
O

0)

I

11

X>O5 OO *> (M UT5 <N

l i 1

JS

1

a

2

a

00

V

N

J^-p

*** 0^

1

i

S fl

o

S i

H

o

^ ? * CO O ^* ^O

fcoa

i I O O ^^ ( ^

RESULTS OF SALMON MARKING 67

In sixty-nine Scottish and forty-eight Irish re-
turns of kelts recaptured as clean fish in fresh water
we have therefore in Scotland practically twice as
many long period fish as short period fish, while in
Ireland we have fully twice as many short period
fish as long period fish two somewhat opposite con-
ditions at first sight. The Irish kelt weights are on
an average considerably less than the weights of the
Scottish fish, this being largely owing to the heavy
Tay kelts in Scotland, but any idea as to there being
a difference in habit amongst light and presumably
younger fish seems dissipated by the middle group
of Scottish returns, where in 1902 and 1903 we
have out of a total of fourteen recaptures seven short
and seven long period fish. This only exemplifies
what has already been said as to weight of fish in
particular localities being no guide for comparison.
The interesting point to notice, however, is that
while in the first group of Scottish and Irish fish
there is an insignificant appearance of long period
fish in Ireland, there is in the 1903-1905 group a
marked increase in the proportion of these fish, and
that similarly the proportion of long period fish in
Scotland greatly increases at the same time, there
being in fact an entire absence of short period fish
in this group. In other words, it would appear
that in the years 1903, 1904, and 1905 there has
been some influence at work which has apparently
induced a larger proportion of fish, both in Scotland
and Ireland, to remain in the sea till the end of the
season's netting. The short absence is still the
most common in Ireland, but the proportion of long

68 THE LIFE OF THE SALMON

period fish is considerably greater than appeared in
the earlier years of the Irish marking. In Scotland,
comparing the first and third groups (the second
group is practically a one river group, and therefore
in this connection of less value), we have first an
almost even balance of long and short absence,
and then the long absence becomes completely
ascendent.

It is of little avail to speculate as to the nature of
the influence which has apparently been at work.
The salmon has latterly remained at home in the
sea, be it because he required longer time to feed up
through a scarcity of food, or a greater delight in an
abundance of food, or lack of inducement to ascend
the rivers, we cannot say, but it is right to notice
that in the Helmsdale and Brora, from which the
majority of returns in the 1903-1905 group come,
the chief fishing takes place in spring, the season
when long period fish are for the most part captured.
The condition displayed by these records can scarcely
be the result of accident. The Scottish records
cover a period of eleven years, and the Irish records
a period of seven years. With regard to the average
intervals of time in the two countries, there is little
difference. The short period in Scotland averages
almost exactly five months, in Ireland not quite
six months. The long period in Scotland is fully
fifteen and three-quarter months, in Ireland a little
under fifteen months.

When the average weights of kelts are regarded,
it is apparent that in Scotland or Ireland no
difference is noticeable either in the fish which

RESULTS OF SALMON MARKING 69

have adopted the short period or those which have
adopted the long period. The average kelt in Scot-
land is, however, about 2 Ib. heavier than the average
kelt of the Irish records, the weights being respectively
7f Ib. and practically 5|- Ib. If individual districts
in Scotland were compared, it could be shown that
the same variation exists between kelts of say the
Helmsdale and the Tay, the former being, so far as
these records show, 6f Ib. and the latter 9 Ib. ; but
the purpose of the above tables being to elucidate
the divided migration, the fact that such differences
exist in the one country only serves to bring out
more forcibly the fact that this divided migration
is independent of the weight of the fish. The sub-
ject of the increase of weight in kelt to clean condition
will have to be dealt with more fully later, but to
complete the statement as to the tables the increase
may be briefly given here. The Scottish returns
show an average increase in the short period of
6 Ib., in the long period of 10 Ib. ; the Irish records,
in the short period 4^ Ib., in the long period
lOf Ib.

In the previous chapters the life of the smolt and
grilse has been traced, and evidence has been educed
to show that the smolts which descend in May at
the age of fully two years make their first reappear-
ance in fresh water at two separate seasons and in
two stages of life, viz., in a year as grilse, and in a
year and a half, having spent two winters in the
sea, as small spring fish. This in the life of the
salmon is the commencement of the divided migra-
tion. The tables just given prove the continuance

70 THE LIFE OF THE SALMON

of the habit, and show that fish of all weights adopt
at one time or other this varied period of marine
sojourn. It will be instructive now, however, to
refer to a few concrete cases by way of illustration,
showing by the actual data collected respecting
individual fish the nature of the evidence, for tables
such as these often conceal truth as language con-
ceals thought. And first we may conveniently
take the case of grilse, and see what becomes of them
after spawning and descending to the sea.

Amongst the Tay records we have two marked
grilse kelts which may be selected. They were
numbered 9402 and 8044. The former adopted the
short period in the sea, and the latter the long
period. No. 9402 was recaptured after five and a half
months (the average period) as a summer fish weigh-
ing 10J- Ib. The record of marking and recapture
is as follows :

9402 / 4 lb- 25 " Kelt S 1 " 1186 Fem< Feb ' 5 1903 Tay U PP er water
\ 10 30" Clean salmon Fern. July31,1903 Tay Estuary nets

The other grilse kelt (although we may observe that
the length is remarkable for a grilse), after spending
fourteen months in the sea, or an autumn and winter
longer than the first-mentioned fish, returned as a
clean spring fish of the large class, a class, be it
observed, which is more pronounced and definitely
constant in the Tay than in any other river in Scot-
land. This record is as follows :

8044/ 6 i lb - 29" Kelt grilse Fern. Jan. 9,1902 Tay Stanley
\19 36" Clean salmon Mar. 14, 1903 Tay Benchill

RESULTS OF SALMON MARKING 71

Other records exactly analogous to these have
been obtained from the Tay marking.

The next stage which has to be referred to is that
of the small spring fish, the fish of the same age as
the grilse kelt, but which has not entered fresh water
as a grilse. This class of fish is well marked in the
Helmsdale and Brora, from which a number of records
are seen. It is also well marked in the Dee. As a
kelt it is usually 5 or 6 Ib. in weight. A Brora fish
marked 7303 is an example of short period migration,
and another Brora fish, 6040, maybe taken as show-
ing the long period. The former returns in three and
a half months as a clean fish of 9 Ib., the latter returns
in fourteen months as a clean fish of 16 Ib.

The records are :

SHORT PERIOD.

f5 Ib. 31" Kelt Fern. Apr. 16, 1901 Loch Brora
7 \9 (?) Clean Fern. Aug. 2, 1901 Kiver Brora

LONG PERIOD.

i

6 Ib. 27" Kelt Fern. Mar. 15, 1901 Loch Brora

6040 116 33" Clean Fern. May 15, 1902 Loch Brora

This completes in a manner the grilse and its
prototype the small spring fish, but it may be asked
whether grilse kelts which are subsequently found
as summer salmon continue to be annual spawners,
or whether one fish may be at one time an annual
breeder and at another a fish of long absence in
the sea a long period feeder. Two other Brora
records which show remarkably small increase of
weight in a two years' interval seem to indicate that
in both cases the greater part of the time between

72 THE LIFE OF THE SALMON

marking and recapture must have been spent in fresh
water, and that annual spawning had prevented any
very substantial increase.

In both records the fish were recaptured as well
as marked when in the kelt condition.

The records are :

28" Kelt Fern. Mar. 15, 1902 Loch Brora

34" Kelt Fern. Mar. 28, 1904 Loch Brora

/4 25" Kelt Fern. Apr. 20, 1901 Loch Brora, a grilse

7298i

/

16 J 30i" Kelt Fern. Mar. 18, 1903 Loch Brora

Only a study of the fish's scales could yield
definite evidence, but, arguing by analogy, the pre-
sumption is that both fish had before the date of their
recapture spawned three times. The weights offer
strong contrast to the case of the clean spring fish
of the Tay which has only reproduced its species
once, but has, although apparently a year younger
than the 4 Ib. Brora fish, already reached the weight
of 1 9 Ib. This is an example of extreme difference
between fish continuing in the short period habit
and a fish, from a river remarkable for its large fish,
after spending the long period in the sea. No such
difference could be found if the more natural com-
parison be made of short period with short period
fish and long period with long period fish.

In continuing the progressive examination we
have now to deal with the kelts of the summer fish,
which, after having spawned as grilse, appear on
their second annual ascent as fish of 10-14 or 15 Ib.
in a river such as the Tay. In spawning a female
salmon loses approximately one-fifth of its weight.

RESULTS OF SALMON MARKING 73

We therefore select Tay kelts weighing 8-12 Ib.
In doing so we are still, I consider, fairly free from
the possibility of including fish of another class, so
long as we restrict ourselves to returns from one
river ; and in making such restriction preference must
be given to the'Tay, not only because the classes of
fish are there well defined, but also because the
numbers of returns are considerable. We are still,
of course, dealing exclusively with kelt to clean
fish increases. A fish showing the short period in
this class is No. 8171. It was seven months
between marking and recapture, and from a kelt of
11 Ib. had become a summer fish of 17 Ib. The
record is :

f 11 Ib. 38" Kelt Fern. Jan. 18, 1902 Tay Battleby
8171-5 1PT Clean Fern. Aug. 20, 1902 Tay Estuary nets

flllb. 38"
117 -

As an example of very rapid growth in the Tay
we have No. 7780, a kelt of 12j- Ib., which doubled
its weight in six months. With regard to the fish
showing the long period, it may be said that, if they
are of the same class as the fish just referred to, and
have spawned in two successive years, their future
record shows that a salmon may at one time of its
life adopt the short absence in the sea and afterwards
stay for the long period. Five fish could be so cited
which from being kelts of 8-12 Ib. have become fish
of from 19 to 25 Ib. in from a year to a year and
a half. One record may suffice :

/ 8 Ib. 38" Kelt Fern. Feb. 6, 1902 Tay Battleby
)8 \20-3 38" Clean Fern. Aug. 18, 1903 Tay Estuary nets

In all these stages we see this same divided

74 THE LIFE OF THE SALMON

migration, and without any doubt the same habit
subsists throughout the life of the salmon. When,
however, we reach the stage in advance of that
just presented, it becomes impossible without the
assistance of scale examination in each case to
identify with certainty the particular class of fish
marked. The weights of the summer fish last re-
ferred to are from 19 to 25 lb., but the weights of
the large Tay spring fish are, we have already seen,
very similar. Therefore the arbitrary selection of
kelts is almost certain to include both classes of fish.
To carry out the same system of selection is, more-
over, to deal with really heavy fish, which, as every
one who has long looked forward to the landing of
a thirty- pounder knows, are not to be met with
every day. Fish much heavier than 30 lb. are of
course taken from time to time, but we are approach-
ing the normal limits of the salmon's growth in
Scotland or Ireland, and heavy kelts which are
marked do not for the most part reappear. I can,
however, instance two Tay fish and one Deveron
fish which take us beyond 30 lb. The Tay fish
show the long period in the sea, and the Deveron
fish shows the short period, so that we may fairly
assume that the twofold habit obtains till a late
period in the salmon's life. The records are :

ri41b. 36" Kelt Fern. Jan. 17, 1902 Tay
3l

ri41b.
\33

43" Clean Fern. July 27, 19C3 Tay

/161b. 36" Kelt Fern. Feb. 11, 1904 Tay

6 B \34 43" Clean Fern. July 17, 1905 Tay

f 19 lb. 38" Kelt Fern. Mar. 4, 1902 Deveron

6541 \34 41" Unspawned Fern. Oct. 22, 1902 Deveron

RESULTS OF SALMON MARKING 75

The Deveron kelt was reported as very well mended
when marked at Muiresk. Its condition when re-
captured at Ardmiddle, which is just above Muiresk,
and in the upper waters, clearly showed that the
fish would soon have spawned, i.e., within about
eight months of the date of marking.

The data dealing with the kelt to clean fish con-
dition form only one section of the salmon marking
returns, but introduce the particular migratory
movements, which are perhaps of greatest practical
importance to our salmon fisheries. The dual habit
of migration shows us that a large number of salmon
remain in the sea even when their fellows are re-
producing their species in the rivers. So far as the
returns show, the proportion of what may be termed
the feeding fish is subject to variation in different
localities and in different seasons, and is much
greater in Scotland than in Ireland. Moreover,
we do not at any time either in our rivers or in oar
sea salmon nets recover a large proportion of the fish
marked. The marking experiments conducted in
connection with the North Sea Fishery Investiga-
tions, in which large numbers of flat-fish have been
marked, have yielded a much higher percentage of
returns than have ever been obtained in marking
of salmon. This is rather striking when it is recol-
lected that the salmon congregate in the confined
waters of rivers, and are caught in such localities in
large numbers. Some do no doubt lose their marks,
but we have no evidence to show that this happens
often. Marked fish may be recaptured and not
reported, but with the interest now taken in the

76 THE LIFE OF THE SALMON

matter and the wide circulation of notices respecting
marking, and the offer of rewards, I do not think
it likely that many records are lost in this way.
Further, the examination of salmon scales seems to
indicate that some fish may pass over two or it may
be three consecutive spawning seasons.

So far as the Scottish returns go, it is clear that
the majority offish spend much more time in the
sea than in fresh water ; and the Irish returns,
although fish showing the short period are in the
majority, and their average absence about six months,
nevertheless yield a proportion of long period fish,
so that here also most time is spent in the sea. It
is quite conceivable, therefore, that with our netting
in the mouths of rivers and close in shore only on
the coast at a definite season we fail to trace a
proportion of our marked fish. Yet in contradis-
tinction from this it has happened that from some
36 kelts caught by rod and marked in the Upper Tay
alone, as many as 22 per cent, have been recaptured.

The stock of salmon is, however, entirely depen-
dent upon the fish which enter the rivers, since
salmon do not spawn in the sea and the ova are incap-
able of development if shed in the sea water. Hence
the importance of moderation in our netting of rivers
and estuaries of rivers. It is, of course, clear that
amongst netted rivers those of large volume with
wide estuaries are more capable of maintaining a
suitable stock of breeding fish than small rivers with
narrow mouths, since in the former a proportion of
each run of fish is more likely to pass the nets,
provided the gauntlet of netting is not too long to

77

render the weekly close time inoperative. But it
should be borne in mind that a proportion of each
run of fish is essential to the proper maintenance of
a suitable breeding stock in any river. In operating
upon the stock of salmon in the sea, however, we
are drawing upon a larger supply ; we are not catch-
ing definite runs of fish which are going to breed in
the near future, but are catching fish which are
moving off and on, or along, the coast. Some of
them may be on their way to a river, many of them
are not, but are fish living in the sea. The influence
of such fishing upon the breeding stock is small as
compared with the influence of the river net, and the
fish so caught and marketed are in the finest con-
dition possible. From statistics showing the results
of different methods of regulating salmon fishing, as
practised in the past not only in Scotland but also
in Norway, there is, moreover, ample proof that the
sea rather than the river is the place for the net.

I am aware that in England this view is regarded
by many as the rankest heresy. The English Act
of 1861 practically abolished coast netting in that
country, but permitted the pernicious policy of fish-
ing rivers not only by nets but by various forms of
fixed engine. A complete reversal of this policy has
been found efficacious in restoring depleted salmon
fisheries elsewhere (when other contributing agencies
have not also to be dealt with) and is, by all argu-
ments in the case, the sound course to follow. To
capture 15,000 to 18,000 salmon on the coast of a
single Scottish district, where the breeding stock is
maintained by a well-stocked river now all but

78 THE LIFE OF THE SALMON

cleared of nets and famous for salmon angling,
should be sufficient to show that such a policy re-
sults in the creation of a valuable sea fishery. In
the river in question, moreover, between 5000 and
6000 fish can be taken by sweep-net at the river
mouth without interfering with the full supply of
fish for angling and breeding.

A question most commonly asked when salmon
marking is mentioned is, Do salmon return to their
own rivers ? The widespread belief that they do is
in ninety-nine cases out of a hundred based on mere
hearsay evidence, but it is none the less satisfactory
to know that the salmon marking experiments verify
the belief. As may almost already have been
gathered from the cases cited, the great majority
of fish do so return. From time to time, however,
fish turn up in other rivers as exceptions to the
general rule. In Scotland a Helmsdale fish was
caught in the Halladale, 90 miles distant by the
coast line. Similarly a Spey fish was caught in the
Dee, 90 miles distant. The first fish went north,
the other south. Other transferences are less dis-
tant. Helmsdale to Brora, a distance of only 12
miles Deveron to Spey, and Spey to Deveron. In
Ireland a remarkable case is reported by Mr. Holt
in the first report of the Irish marking. The fish
was marked D 95, on January 1, 1901, after having
been caught, conveyed to Kilrea, on the Bann, and
impounded there for hatchery purposes, stripped,
and then allowed to rest for three days. The re-
capture was made 200 miles distant in the tidal
waters of the Bundrowes, on April 9, 1901. The

RESULTS OF SALMON MARKING 79

fish appeared still to be a slat or kelt, and showed
no increase in weight (7 lb.). In view of the rather
peculiar nature of the estuary in which the fish was
found, Mr. Holt suggests that possibly the fish had
no intention of leaving the sea for fresh water, but
having found its way into the basin of the river
mouth during spring tides, had found its retreat cut
off. In view of this explanation, the record is similar
to a - privately marked Scottish fish which travelled
from the Grimersta on the west side of the island of
Lewis to Castletown, which is on the coast east of
Thurso, 145 miles distant. The fish was marked by
Mr. Byres-Leek, "J. B.-L. 10," in March 1902,
when a kelt. Five months afterwards, i.e., in
August, when recaptured at Castletown, it was still
a kelt. I have the specimen preserved, as an August
kelt is in my experience unique, and this fish showed
not only the lank condition, but the abrasions below
the jaws and on the ventral surface of the body, so
commonly associated with the recently spent fish.
Like D 95, the fish was a male, and it seems pro-
bable that the kelt condition may have subsisted
through the fish having been originally perhaps only
half spent, and having afterwards entered some river
on the north coast of Scotland, such as the Naver,
Polla, Halladale, or Thurso, in its journey along the
Pentland Firth, and then only at a later date com-
pleted the shedding of its milt. There is, I think,
a fair amount of evidence to suggest that under the
most ordinary conditions the male fish which we see
paired with females at spawning time do not com-
plete the shedding of their milt, or do not always do

80 THE LIFE OF THE SALMON

so (unless the female is a large one) in fertilising the
ova of one female, but, like blackcock among birds, are
somewhat catholic in their temporary attachments.
That a partially spent or completely spent kelt may
on occasion descend to the sea and ascend another
river has been evidenced on more than one occasion.
The Spey to Dee fish already referred to is an in-
stance ; another Spey fish a male of 30^- lb.
after being handled at the Fochabers hatchery,
turned up again in the Deveron in three months as
a kelt, believed to be not yet completely spent,
weighing 28 lb.

It is rather striking that the fish which show
those pecular reascents are male fish which have
been a good deal handled and stripped at hatcheries,
and in all probability the descent and reascent would
not have taken place in this apparently unnatural
manner had the fish paired and spawned and pos-
sibly again paired without interference. A marked
example of this peculiarity is seen in a fish which,
like D 95, was conveyed to a pond and retained for
hatchery purposes at Sandside. It was caught in
the sea during the fishing season of 1905, and held
up in fresh water till spawning time. It was
stripped (a male), marked on January 15, 1906, and
turned loose in a small stream not far from the
hatchery. As is well known, no kelt seems to care
to remain in a small stream for any length of time,
and we are therefore not surprised to learn that the
fish went to the sea without delay ; but this fish was
captured up the river Thurso, fourteen miles up, in
sixteen days. A frightened creature commonly and

81

by instinct turns quickly in the direction in which
it has last come where it was not frightened ; a
frightened ascending fish goes down, and a frightened
descending kelt goes up. In a small stream, how-
ever, escape by ascent is scarcely possible, and a
forced departure from the course prompted as natural
must only heighten and complete the alarm and
demoralisation of the creature. In the small streams
of the West Highlands the kelts which quickly and
naturally descend recover condition in the brackish
water at the mouths of the streams. If driven from
such a haven of refuge by fear or any other com-
pelling cause, recovery may be sought in fresh water
elsewhere.

It does not always follow that fish caught in the
sea at a considerable distance from their breeding
river will not return thither. Yet in the search for
food and the following of herring shoals it is not
surprising that some fish may lose their bearings.
In the drift-net fishing for salmon which is conducted
off the coast of Ireland, nets appear to be shot at
times as far as eighteen miles from shore, and if fish
wander, say double or treble this distance, as salmon
marking shows us they sometimes do, they may
naturally resort to other rivers than were formerly
visited ; or, as has been suggested, some fish may
systematically wander to rivers commonly frequented
by larger fish, rather than return, after growth in
the sea, to the rivers of their youth. In the salmon
marking results we do not, however, find much evi-
dence of this. The numbers caught in bag -nets on
the open coast at a considerable distance from the

F

82 THE LIFE OF THE SALMON

rivers in which they were marked represent only
8f per cent. It is remarkable in this connection
that although a great number of fish have been
marked in the Tay (of which eighty-six have been
recaptured), no Tay fish has been recaptured either
in another river or in coast nets. The great volume
of water reputed the largest in the British Islands
may make itself felt for avery considerable distance
seawards, and so give a guide to Tay fish in the sea.
Certainly to any one who has sailed much off the
coasts of Fife and Forfar the influence of the Tay
when swollen and discoloured by flood is very clearly
seen, stretching as it often does in a great belt of
discoloured brackish water in a southerly and easterly
direction beyond the limits of St. Andrews Bay out
into the open sea.

It is of some interest to notice the direction in
which the Scottish fish caught in sea nets at some
distance from the rivers in which they were marked
chiefly travel. The Grimersta-Castletown fish, cross-
ing the Minch and travelling the whole length of
the Pentland Firth in an easterly direction, has
already been referred to. Six Helmsdale fish
travelled south and have been caught at Portma-
homack, on the coast south of Tarbat Ness. Three
Helmsdale fish went into the Brora, and one north to
Dunbeath ; another north to the river Halladale. A
Brora fish also went north 100 miles, and was caught
in a bag-net near the mouth of the Halladale. The
contrast in the increase of weight in these two other-
wise similar records is very remarkable. The Brora
fish increased 5j Ib, in eighty-eight days, while the

RESULTS OF SALMON MARKING 83

Helmsdale fish took 506 days to increase 10 lb.,
nearly double the increase, but in almost six times as
many days. Probably, however, the second fish had
spawned in the interval. The majority of the coast
recaptures show movement in the opposite direction
from those two similar records. Five fish have
followed a very similar course in going from the
Deveron right round out of the Moray Firth in an
easterly and southerly direction to the neighbour-
hood of Aberdeen. The intervals between marking
and recapture are 88, 109, 122, 126, and 187 days,
and the increases in weight respectively 4j, 3f , 2f ,
9, and 4j lb. In the last-mentioned fish the
estimate of weight could alone be obtained, as the
mark was not noticed till the fish was displayed in
a fish shop in Newquay, Cornwall. The increase of
9 lb. in the kelt weight in four months is consider-
able, more especially since the fish was a 1 6 lb. male
kelt when marked.

Twenty coast recaptures are charted in the twenty-
fourth Annual Report of the Fishery Board for
Scotland, and of these, fifteen show a southerly and
easterly and only five a movement in the opposite
direction. It is unwise to lay too much stress upon
a result which after all is deduced from a limited
number of recaptures, but the trend of the currents
corresponds with the direction in which the majority
have travelled.

Before completing the records which are obtained
from kelt marking, we have to deal with the fish
which are afterwards recaptured, as kelts. These
fall into two classes : first the fish which have been

84 THE LIFE OF THE SALMON

to sea in the interval and have again ascended and
spawned, and second the kelts which have never left
the fresh water.

In comparing the first class we have the possibility
of noticing, better than at any other time in the fish's
history, the actual net increase of weight, since at
both times of observation the fish is reduced to the
same condition. Clean fish vary in weight according
to the time they have been in fresh water, and the
state of their nutrition or genitalia. In the sea they
are still possibly increasing in weight, and their
record is in this respect incomplete ; in fresh water
they are losing weight. Unfortunately it happens
that recaptures of this kind are scarce, and at the
best rather conflicting. Of course, the catching of
kelts is not pursued in any sense as is the catching
of salmon. Nothing is to be made out of a kelt
when he is caught. I might first mention two
Brora fish which have evidently spawned annually
for three consecutive years :

Ib. 28" Kelt Fern. Mar. 15, 1902 Loch Brora

7225 ^ 3r Kelt Fem Mar 28, 1904 Loch Brora

T4 Ib. 25" Kelt Fem. Apr. 20, 1901 Loch Brora, a grilse
\6i- 30" Kelt Fem. Mar. 18, 1903 Loch Brora, a salmon

In the former the increase is 2|-, in the latter
2 J Ib. ; the interval is practically two years in each
case. Two examples of Deveron fish show better
increment :

12 Ib. 36" Kelt Fem. Feb. 22, 1901 Deveron

6504 U6 38i" Kelt Fem. Mar. 14, 1902 Deveron

|16 Ib. 37" Kelt Fem. Feb. 13, 1903 Deveron, Muiresk
6560 \22 42" Kelt Fem. March 1904 Deveron, found dead

1 mile above mouth

RESULTS OF SALMON MARKING 85

In the former the increase is 4, in the latter 6^ lb.,
and in each case the interval is thirteen months.

Another record is from the Tay, but as the fish
was found dead, and the weight merely estimated,
the result an increase of 3 lb. in exactly two years
is not so valuable as otherwise would have been
the case. So far as it goes, however, the record
corresponds with the slight increase seen in the Brora
fish mentioned. The last record which I can mention
from Scotland is rather remarkable, since in a year
all but ten days the fish is reported as having
gained only \ lb. in weight and nothing in length :

/13 lb. 38" Kelt Fern. Mar. 15, 1902 Loch Brora
7355 \13 38" Kelt Fern. Mar. 5, 1903 Loch Brora

All these fish are clearly annual spawners, and
since long period migrants are, in Scotland, in the
majority and remain in the sea, their recapture in
the kelt condition is less likely to occur. The small
increase even in annual spawners is sufficiently re-
markable. There are no available records from
Ireland or England.

With regard to the manner in which kelts
descend to the sea, and more especially the time
they occupy in such descent, it may be useful
first to recall the experiences of a former super-
intendent of Tweed bailiffs, as stated at p. 66 of
the Tweed Reports, 1866. Mitchell relates that in
1854 he marked in the Whitadder 206 bull-trout
kelts and took them down to the tideway at or near
Berwick Bridge. " Few, if any, of these went down
to the sea, as for many weeks after they were

86 THE LIFE OF THE SALMON

occasionally caught in the Whitadder, and a number
of them were got up the river as high as Norham."
In the Tweed Report for 1875 it is stated: "The
Committee met at Melrose to ascertain if it is prac-
ticable, when a great number of salmon kelts were
collected in the cauld pool there on account of dry
weather, to catch them with a net and put them over
the cauld. A few shots were rowed and eight fish
caught ; but the experiment was not successful, as
the fish put below the cauld did not go down the
river, but remained in shallow water within reach of
poachers." In the Norwegian Fisheries Report for
1895-96 Herr Landmark tells of a similar experience
on the Aensira river, where two waterfalls exist. At
both waterfalls fish-passes have been erected, and
the lower pass was used as a convenient place to
capture fish for hatchery purposes. The pass was
therefore blocked by means of a grating inserted in
its upper end. Fish, having been stripped, were
placed above the fall, yet many of these fish were
afterwards found in the fish-pass, having apparently
descended by the fall and re-ascended in the pass.
One such fish was recognised as doing so five times,
although on each occasion after recapture it was
put into the river farther and farther up stream.
The seaward migration of kelts in Scotland is of
course not influenced by the complete freezing up of
parts of rivers, as is the case in Norway, where it is
often, I understand, physically impossible for kelts
to descend till released by the thawing of the ice.

With us, however, the descent in large rivers ap-
pears to be much slower than in small rivers. Kelts

RESULTS OF SALMON MARKING 87

can scarcely be caught in the small streams of the
West Highlands, so quickly do they get away to the
sea. In a large river, on the other hand, or in any
river which broadens out into a large lake or into
deep pools, the descent of female kelts at least is
very slow indeed. A large number of recaptures of
such descending kelts have been made in the Spey
and Tay, and it would appear that so long a time as
three months may be occupied in descending forty
miles of river. Grilse kelts, as already stated, descend
to the sea with some promptitude, but as salmon
grow older their descent after spawning becomes
slower, although this appears to be less true of males
than of females. Large female kelts linger longest.
The discovery of this habit makes a considerable
difference in our interpretation of the increments
put on in the sea. One fish may relatively be so
much longer in reaching the sea than another even
if both be, say, short period fish. In judging of in-
crease from kelt weight this uncertainty of knowing
when the sea was entered is always present. We
may have in long period fish considerable uniformity
of increase, since an actual delay at starting may in
the longer period be made up. For instance, in
Helmsdale long period fish of 480, 498, 504, and 506
days we find increments respectively of 9, 9, 9, and

10 Ib. Yet clearly there may also be great variation.
In short period fish we have similar results. Two
Brora fish were taken, the one in 162 and the other
in 163 days : the one had added 4 Ib., and the other

11 Ib. Many Tay fish recaptured are described as
" bull-trout." In the Tay a bull-trout is in reality a

88 THE LIFE OF THE SALMON

salmon (S. solar) which is more or less spotted about
the operculum and shoulder, has maggots in its gills,
and which when cut up is found to be pale in the
flesh. It is not the bull-trout of the Tweed and
Coguet, otherwise called the round- tail. These
Tay bull-trout are not at all uncommon as fish
of over 40 lb., and the natural presumption seems
to be that they have their peculiar appearance
through having adopted a more estuarial habit and
different feeding than is common with normal S.
solar. From Tay salmon which have been described
as bull-trout on recapture eight examples may be
selected. Four of them show the short period, and
four the long period ; two were described as un-
spawned grilse on marking. The particulars are of
some interest, and are as follow : *

SHORT PERIOD.

A 6 lb. Kelt after 196 days increased 6 lb.

4 )) )> l " 11 "i 11

7f Unspawned Grilse 265 4

j) 5-J 11 11 ^"^ 11 11 ^ 11

LONG PERIOD.

A 14 lb. Kelt after 556 days increased 19 lb.

*" )> i) 447 ., ,, 4 ,,

556 8J ,,

1^ 525 ,, ,, lo

Here the uniformity is seen to be in the short
period fish and a very marked extreme in the long
period fish, but we must recollect that possibly the

* " Arretor," Proc. Eoy. Soc., Edin., XXV. Part I. p. 27.

RESULTS OF SALMON MARKING 89

second fish may have spawned in the interval of
447 days, which is, of course, the interval between
marking and recapture, not necessarily an indication
of the length of time spent in the sea. Yet even
without the second fish we have the first and third
each with a precisely similar interval, yet one with
twice the increase of the other. That these bull-
trout do go to the sea may be inferred from the fact
that sea lice are commonly found upon them, as well
as gill maggots, when they come into the river.
They are by no means confined to the Tay, but occur
sparingly in other parts of the country. Personally
I have noticed them also in the Ness, Helmsdale,
and Dee. Gtinther has described them from the
Beauly, and Parnell from the Forth.

The last series of observations on the habits of the
salmon to which reference may here be made has
been obtained by the marking and recapture of clean
run fish. Now the habits of clean run salmon,
otherwise spring fish, have given rise to much dis-
cussion, very largely because some rivers contain
spring fish, while others, for no apparent reason, do
not. It is proposed in another chapter to deal with
early and late rivers, and in this rather to deal with
the habits of the spring fish after it has entered its
" spring river." At the outset, however, it seems
necessary to recollect that the habits of spring fish
seem to be subject to considerable variation under
the conditions which obtain in different countries.
In Norway, Iceland, Lapland, and possibly elsewhere
clean summer fish ascend freely in snow water. In
Scotland " snaw bree " prevents fish running ; by

90 THE LIFE OF THE SALMON

waiting they can, I presume, secure passage in a much
more agreeable fluid. But in the normal early river
the early fish have already advanced well up stream
before winter descends upon the hills. If they
encounter floods of cold, snowy water they may drop
back to the warmer estuary if they have not ascended
to upper waters, but in Scotland we have not evi-
dence that clean fish are in any sense temporary
visitors to fresh water or habitually drop back into
the sea. In Ireland the evidence secured by mark-
ing clean fish seems to point to a rather different
conclusion, as described by Mr. Holt.* Out of
thirteen recaptured Irish clean fish there is evidence
that three were in the sea in the interval. These
fish were numbered 433 D, 858 D, and 861 D, and
the last two were presumably feeding in the sea,
since each on recapture showed an increase in weight,
858 of 2|- Ib. with an interval of thirty-seven days,
861 an estimated increase of 1 or 2 Ib. and an
interval of thirty-nine days. Mr. Holt further states
that this is in conformity with the belief held at Lis-
more that these winter fish make only a temporary
stay in fresh water, and are by those who catch them
commonly called "droppers," from their habit of
dropping down stream. This view is moreover
strengthened by the fact that 858 Dwas recaptured
in the sea, in Youghal harbour, about eighteen miles
below Lismore weir in the Blackwater, where the
fish was marked. I am not aware that anywhere in
Scotland the view is held that clean winter fish are
only temporary visitors to fresh water.

* Report for 1901, Part II., " Scientific Investigations," p. 195.

RESULTS OP SALMON MARKING 91

That such fish may drop down stream after making
a comparatively short ascent and meeting floods, as
stated, is believed by netsmen, since after cold winter
floods they not infrequently catch coloured fish in
the lower reaches and estuaries ; but it also happens
that occasionally such fish are found to be affected
by disease, and to be therefore in a weakened con-
dition. The early running Tay fish ascend to Loch
Tay about fifty miles, and there yield regular sport
till May, when they begin to ascend still farther.
The early Ness fish run quickly through the six
miles of the river, and in the same way yield sport
in Loch Ness, from which at an early date they also
ascend the Oich to Loch Oich and the Inverness-
shire Garry. These upper waters are well stocked
with fish on the opening day of the rod fishing,
February 11, and continue to be so. The same may
be said of rivers farther north, the Thurso and
Loch More at its head, or the Naver with Loch
Naver at its head. In all such localities there seems
no indication that the fish, once well up the rivers,
descend again. On the contrary, the indications all
seem to point to the conclusion that these fish remain
in fresh waters during the whole of the ensuing
fishing season.

In the Spey 150 clean fish were marked during
the close time of 1896-97, and of this number
67 were recaptured after the opening of the
fishing season. The lower reaches of the Spey are
very rapid, and the winter in question was excep-
tionally severe, so that one would expect ascent to
be slow and that not a few would descend these

92

lower reaches to some extent. As a matter of fact,
34 had moved up, 25 had dropped lower, and the 8
remaining fish were found in the position in which
they had been marked. Only 1 fish out of the 25
which had dropped back was reported as having in the
interval of thirty-eight days gained ^ Ib. in weight,
and one naturally hesitates to state that with so slight
a change one has evidence of sea feeding and not
rather of inaccurate weighing. All others of this
series showed decreases from to 2 Ib. In the Tay
24 clean marked fish (spring fish) have been recap-
tured within short intervals. They were marked
between Perth and the Linn of Campsie, about nine
miles above Perth. Of this number 8 fish which
had dropped back were diseased, with one exception,
No. 9100. This fish was marked at the Linn, and
recaptured in a quite healthy condition about 13
miles farther down in the upper tidal waters of
the estuary. The interval was two months and ten
days, and the fish was reported as having lost 5 Ib.
in weight (the greatest decrease recorded for a clean
fish in fresh water 27 Ib. to 22 Ib.). In the interval
a heavy flood of snow water had occurred. Two fish
were found in the same position as before, and two
had ascended between 30 and 40 miles, while others
had made a less ascent. A few recent Dee recap-
tures of clean fish show in every case an ascent, and
in all but one instance an ascent to the top waters
of the river. In these Scottish records there seems
nothing to indicate any return to the sea, as has been
shown by Mr. Holt to occur in Ireland.

CHAPTER IV

THE SCALES OF THE SALMON AS RECORDS
OF THE SALMON'S LIFE

Age of salmon to be told from their scales Progressive
growth of the scales Lines of growth on scales The
spawning mark Records of marked smolts as shown on scales
Records of small spring fish marked as smolts Infrequency
of the spawning mark Scales of kelt recaptured as clean
fish The average number of lines added Scales of large
salmon Salmon disease The bacillus of origin Fungus
spreads quickly on dead fish

DURING the last three years or so we have been
gradually and with increasing certainty realising
that a study of the scales of the salmon yields a
most valuable addition to our knowledge of the
fish's life.

We have long been able to tell the age of a horse
by looking at its teeth. In like manner we can now
tell the age of a salmon by looking at its scales.
Other hard tissues, such as the bones, would no
doubt show corresponding information the ear
bones or otoliths have already been used for the
purpose with regard to many species but in the
case of the salmon a properly selected scale is by
far the most convenient tablet to consult and is at
the same time a very satisfactory one.

94 THE LIFE OF THE SALMON

When a female fish spawns she rapidly loses 23
per cent, of her weight and a corresponding diminu-
tion of body girth. As a kelt the descent to the
sea is then made, and thereafter, with a sudden
accession of feeding, a marked growth begins, so
that the fish may soon be double its kelt weight.
If this greatly increased bulk is to be provided
with a scale covering, it is clear that one of
two things must happen : either new scales must
grow amongst the old ones, or the already existing
scales must become enlarged must, in fact, have
the power of growth which the body of the fish
possesses.

It has been supposed by some that a shedding of
scales takes place at certain periods of the fish's life,
and that new scales are formed. It has even been
thought by some that the male salmon at the breed-
ing season has not a covering of scales. It is true
that at this time the male fish is remarkably smooth
and slippery, and that the surface is not imbricated
in any way by scale edges. This is not, however,
from any loss of scales, but because the whole con-
dition of the skin has altered and become thickened
and toughened. The scales are now deeply em-
bedded in their skin pockets, and cannot be got
hold of properly unless the thick skin covering is
lifted or punctured. But at no time of the fish's
life are young growing scales found amongst the
existing scales, except in places where repair after
injury is going on. There is no arrangement of
growing rows of scales ready to take the place of
the older scales, as teeth do in the jaws of a Green-

THE SCALES OF SALMON 95

land shark. Moreover, salmon, young and old,
small and great, are found with the same number of
scales when certain rows are counted, and this
character is so constant as to be one of the most
trustworthy for the identification of salmon from
other species.

On the other hand, the progressive growth of the
scale as viewed in fishes of different age, the increase
of size and the definite system upon which this in-
crease of size is clearly brought about, the impress
of the young condition still visible on the old fish's
scale, and the comparison of scales taken at different
times from salmon which have been marked and
afterwards recaptured, all impress the mind with the
conviction that not only does the scale grow as the
fish grows, but that the periodic nature of the
salmon's growth is stamped upon the scales.

In early life the scales are merely circular patches
on the integument of the post-larval fish. These
centres of growth gradually have more and more
scale matter added to their peripheries, till inter-
vening spaces are all covered and the scales com-
mence to overlap. The young salmon is barely two
inches long when the growing scales begin to over-
lap. The scale matter added to the growing sur-
faces appears in a succession of ridges which under
the microscope give the appearance of concentric
lines round each centre of growth, the whole being
covered by a membrane or outer skin.

By the time the smolt is ready to depart to the sea
the overlapping portions of the scales have become
less distinct in their line formation than the portion

96 THE LIFE OF THE SALMON

closely associated with the under skin in the skin
pockets in which all scales lie. In all subsequent
appearances it is this covered up portion of the scale
which chiefly supplies the records of growth, of
migration, and of spawning. The exposed or posterior
area of the scale which is a small portion of the
whole becomes, as life goes on, a mere irregular half-
moon shaped area of comparatively, featureless surface.

For our information respecting the scales of the
salmon and their interpretation we are indebted to
Mr. H. W. Johnston, who in two papers to the
Reports of the Fishery Board for Scotland * has dealt
with the subject exhaustively.

The experiments made at Fochabers on the rear-
ing of smolts have constantly shown that, even
amongst those young fish, food is freely taken during
spring and summer, but much less freely taken at
other times of the year. Dr. Hoffbauer, who studied
the scales of a German carp in 1899, and Mr. Stuart
Thomson, who wrote on the periodic growth of
cod scales in 1902 and 1904, have each shown that
this habit is common to other fishes, and that there
is a distinct result in the appearance of the scales :
that while summer feeding and growth is in progress
the lines or ridges on the scales are added in greater
numbers than at other times, and that between the
ridges the spaces are greater, and that while the
more moderate feeding of winter is in progress the

* H. W. Johnston, " The Scales of Tay Salmon as Indicative of
Age, Growth, and Spawning Habit." Twenty-third Annual Report,
Fishery Board for Scotland, Part II. Appendix II. Ibid. " The
Scales of Salmon." Twenty-fifth Annual Report, Fishery Board
for Scotland, Part II. Appendix II.

9?

ridges are few and close together. The result is that
each summer and each winter leaves its indelible
trace on the scale.

In the case of the salmon the same is true, and in
addition there are periods when, through the for-
saking of the sea for fresh water, where no regular
nourishment of the body takes place, a complete
cessation of scale growth results. Also, when
spawning has taken place, with its consequent
shrinkage of the body girth in the kelt, a record
called the " spawning mark " is stamped upon the
scale, due in part to the apparent fraying of the
scale edges and in part to the healing or mending
of their frayed edges when feeding is recommenced.
The spawning mark is therefore analogous to the
" scar tissue " of pathologists.

But there is in all cases a certain irregularity in
the amount of growth, due directly to the varied
amount of feeding. Even parr and smolts show
differences in the number of ridges added at similar
times. A parr growing in a river which flows through
richly wooded country, and which has water of a
comparatively high temperature, receives many more
chances of food than the parr growing in the rocky
stream of some cold mountainous district. Never-
theless, the number of ridges in each period of
life is found to be the safest guide to the under-
standing of the growth, the separation of summer
and winter areas being at the same time clearly kept
in view.

In Mr. Johnston's second paper (loc. cit.) several
tables are given showing the number of lines of

Q

98 THE LIFE OF THE SALMON

growth in parr and smolt scales, and from these it
is seen that when the smolt is fully two years old
and has reached tidal water, preparatory to the
first migration to the sea, the number varies from
23 to 28. The scale is well defined, and bears upon
it the traces which, as we shall presently see, remain
in all older scale appearances.

It is of special interest to note that in connection
with the smolt-marking experiments in the Tay, to
which reference has already been made, evidence as
to scale growth was collected, and that when those
wired smolts again turned up as grilse, and subse-
quently also as small spring fish, the record of the
scales was obtainable. It may be said at once that
this record, as interpreted by previous study, was
in complete harmony with the evidence obtained
from the marking.

I subjoin Mr. Johnston's table, which refers to
thirteen of the grilse recaptured in 1906 with smolt
wires in situ :

THE SCALES OF SALMON 99

SOME "WIRED" SMOLTS RECAPTURED AS GRILSE

NUMBER OF LINES ON SCALE.

Date.

Lb.

Up to Smolt Stage
(River.)

After Smolt Stage
(Sea).

Remarks.

Total.

1st
Year.

2nd
Year.

3rd
Year.

Total.

3rd
Year.

4th
Year.

1906.

June 1

2-B

24

8

14

2

35

29

6

si

O

26

4i

25

6

15

4

42

31

11

i >,

28

4J

28

16

12

_ ?

45

32

13

Parr lines

.5

doubtful.

H Be

July 4

5

23

9

12

2

44

30

14

|l

,, 10

4

23

9

11

3

43

30

13

|

> 14

5f

20

8

12

-

43

29

14

ii

i. 14

7

24

9

12

3

47

32

15

si

16

6*

25

7

13

5

48

30

18

-g*

si

Aug. 4

5

22

7

12

3

47

29

18

ii

i?

,. 14

3*

24

10

11

3

45

31

14

Exterior ridges

close.

0) <N

,, 14

6i

22

8

12

2

47

29

18

1|

.. 17

6J

26

7

14

5

47

28

19

Ii

;n

9

22

10

9

3

48

31

17

Ridges broad.

r

" From the marking we are sure that the fish in
the above table were smolts in the spring of 1905,
and that they returned to the river as grilse in the
summer of 1906, weighing from 2^f Ib. in June to
9 Ib. and more in August. From the scales we learn
that all the specimens were a little over two years
old in the spring of 1905, when they were migrating
as smolts ; that all had spent one summer and one
winter in the sea, and before returning had com-

100

THE LIFE OF THE SALMON

menced the feeding of the second summer, which is
represented by six lines in the June capture and by
nineteen rather large lines in one of the August
captures."

In following out the scale record we have also
evidence as to the history of three of the small
spring fish which were, like the grilse just referred
to, marked when smolts in 1905. Here again the
evidence of the scales bears out the deductions arrived
at by the marking. The recaptures were made in
the spring of the present year (1907).

SOME " WIRED " SMOLTS KECAPTURED AS SMALL SPRING SALMON

NUMBER OP LINES ON SCALE.

Date.

Lb.

Up to Smolt Stage
(River).

After
Smolt Stage
(Sea).

Remarks.

Ttl.

1st
Yr.

2nd
Yr.

3rd
Yr.

Ttl.

1st
Yr.

2nd
Yr.

1907.

' a>3

Feb. 18 . .

9^

22

12

10

j

65

28

27

*fe ^*S2

Exterior

5-Sc^f

ridges

&ii

21 ..

9

21

13

9

1

52

29

23

very

"S g"S s

broad.

H -S"

g "3 rt

Mar. 15 . .

101

20

10

10

?

59

30

29

- a H>

" An apparent break in the lines of the second
summer is nearly always visible on the scales, and
looks like a voluntary or compulsory diminution of
feeding for some time. Another point seems evident
from the marking, and that is that these small
springers are returning for the first time, and have
consequently never spawned. This corresponds with

THE SCALES OF SALMON 101

what the scales show, but it must be remembered
that salmon similar in size and appearance that have
spawned during the previous winter as small grilse
are also caught at this season, and it is only by the
spawning mark on the scales and sometimes maggots
in the gills that the difference can be detected,
though it sometimes becomes apparent when the fish
comes to table.

" These small unspawned fish constitute the bulk
of the spring run in the northern rivers, and are
comparatively more important to them than to the
Tay, where larger fish a year or more older come in
during the winter and the spring."

Illustrations of those wired grilse and spring fish
scales are given in Plate V. In each case the
smolt scale will be noticed in the centre, and as the
eye passes from the centre of growth along the long
axis of the scale to the anterior border, will be seen
the large area of growth resulting from the first
summer of feeding in the sea ; then subsequently
the contracted-looking band of small growth, mark-
ing the first winter in the sea. In the grilse scale
a short period of spring feeding in 1906 is repre-
sented by only six lines, after which the fish entered
the river Tay and was killed. In the small spring
fish scale, on the other hand, the second summer in
the sea is seen to have been completed and followed
by a second winter, making an addition of 27 lines,
after which the record of this fish ends like a tale
that is told.

The grilse which enter our rivers spawn in the
autumn. Those fish which do not enter our rivers

102 THE LIFE OF THE SALMON

as grilse but remain in the sea till the grilse stage
is past are of course still maiden fish, and after
entering such a river as the Tay as small spring
salmon will spawn a year later than their fellows.
They merge into the run of 10-16 Ib. summer fish
of the Tay which have completed their two annual
rings. The summer class of fish are, however, apt
to be mixed in a large river like the Tay, as we
have now reached a stage in the salmon's develop-
ment when fish which have previously been in the
river as grilse and which have spawned and de-
scended for the short period of marine sojourn may
be once more in the river. The proportion of such
fish is, however, comparatively small, at least in the
large rivers of Scotland.

The study of the scales has prominently called
attention to the infrequency of spawning amongst
heavy fish. It comes as a surprise, for instance, to
learn that very many of the large spring fish of the
Tay fish almost invariably about 20 Ib. in weight
have never spawned. The scale shows the third
annual ring in course of formation subsequent to
the two-year-old smolt stage ; i.e., they are approxi-
mately five years, yet their scales bear no spawning
mark. They are entering fresh water for the first
time, although many fish of smaller size have
already propagated their species. Such information
makes us realise with additional force the value of
protecting fish when they do breed.

Now let us deal with the records of adult salmon
marked in fresh water as kelts and subsequently
recaptured as clean fish. And first 1 will take the

THE SCALES OF SALMON 103

case of a Helmsdale fish which was marked on
December 12, 1904, when a kelt of 5 Ib. weight.
The number of the mark was 1180 B, and other
particulars respecting it are given in the chapter on
marking (p. 65). The fish was recaptured, clean
and 14 Ib. in weight, on April 30, 1906.

I have included two photographs with which
Mr. Johnston has kindly supplied me, showing first
the scale at time of marking, and second the scale
at time of recapture. Horizontal lines have been
drawn to show the progressive development, and
each stage or interval has been lettered con-
secutively. In the photograph of the scale from
the kelt the history of the fish is at once seen.

A. Parr and Smolt Stage, 1900-02. River.

B. 1st Summer . 1902

C. 1st Winter . 1902

D. 2nd Summer . 1903

E. 2nd Winter . 1903

F. 3rd Summer (till

May or June) 1904 River.

The fish then spawned in the winter of 1904, and was
marked in December as stated.

In the photograph of the scale from the same fish
on its recapture as a clean salmon we have, in
addition to the appearance already seen in the
smaller photograph, first the spawning mark on
recovery from the condition succeeding that shown
in F., and then two additional areas :

G. 5th Summer . . 1905 Sea.
H. 5th Winter . . 1905-6 Sea.

The fish returned to the Helmsdale river in the

Sea.

104 THE LIFE OF THE SALMON

spring of 1906, and was caught on the date already
given.

This Helmsdale fish is typical, and the photo-
graphs show more plainly than any description can
do the characteristic appearance of the kelt and
adult clean fish scales. The magnification is in each
case 12 diameters. The additions G. and H. on
the later scale may appear slight in compari-
son with those seen in the kelt scale, but it is to
be recollected that as the age of a salmon increases
the rate of growth decreases. The first two summers
in the sea invariably show the greatest growth in
the fish.

Evidence from scales of fish which have been
marked and recaptured is evidence of the most
reliable kind, and several instances might be cited
similar to the Helmsdale fish marked 1180. On
account of the nature of the evidence also, I consider
that there need be no hesitation in deducing evidence
from scales of fish for which no marking record is
available, provided always that the observer has a
competent knowledge of the subject and the scale
be taken from a suitable part of the fish, e.g., the
" shoulder."

With regard to the number of lines which will be
found on the scales of fish of different ages, it may
be said that from marked kelts recaptured as clean
fish, a scale having been obtained only at time of
recapture, Mr. Johnston has found that after the
smolt stage is passed a grilse may add 25 to 28 lines
the first year ; if then it enters the river to spawn
as a grilse the lines will vary according as it enters in

THE SCALES OF SALMON 105

early or in late summer or autumn, the variation being
from 7 to 24. lines. If in this second year the fish
remains feeding in the sea the long migration habit
the addition, in the case of Tay fish, may be from 20
to 31 lines. If thereafter the fish enters the river
as a summer fish, very few additional lines may be
added at the commencement of the third year, though
a later fish may show 12 or 15 lines. If in this third
summer the fish still does not enter the river, but
delays till the following winter or spring, 17 to 28
lines may be added in the sea.

From the evidence at command it appears to be
somewhat unusual for a fish to remain till its fourth
sea year without spawning, but a few instances are
on record. Fish which on recapture are from 30 to
35 Ib. show either five or six years' growth on the
scales after the smolt condition, and have spawned
either once or twice. A record of a 42 Ib. fish shows
six years' sea life, or in other words is eight years old
and has spawned twice. The fact that the recapture
was made in fresh water points to the conclusion
that spawning, or rather reproduction of the species
(for the majority of those large fish are males), would
be engaged in for a third and probably last time.

The few records of marked fish which have been
recaptured when 40 Ib. or more than this weight
have, from the information gained from marking
alone, been set down as of eight years old, and in
this connection ib is interesting to recollect that
trout kept in rearing ponds for hatchery purposes
are found to become poor as regards fertility after
their eighth year.

106

If salmon live much longer than eight years they
apparently need not be looked for in the fresh water
spawning places of their species, and the fact that
the majority of the very large salmon captured are
males would seem to indicate that in all probability
the male, as in other animals, remains fertile longer
than the female. So far as I know, the record weight
for a Scottish salmon is 84 lb., but the fish was taken
in the estuary of the Tay, though both in the Dee
and in the Tay fish of over 70 lb. have been taken
in fresh water.

From the study of the scales alone do we gain
this information as to the infrequency of spawning,
and the consequent benefit to our stock of salmon in
preserving most religiously the breeding fish which
enter our rivers. From the study of the scales also
it is possible to note early running fish. Mr. Johnston
has, for instance, found the kelts of spring fish only
a short distance above the tide, as early in the
spawning season as November. Mr. Berrington,
formerly Assistant Secretary to the Board of Trade,
has specially noted in his official reports that when
a river is becoming over-fished the first runs of fish
to disappear are the spring runs. Now the spring
fish are not only of great value in any river, because
they are fish in beautiful condition, which afford the
best sport to the rod, and are the best river-caught
fish for the table, but also because they are the fish
which ascend to the highest tributaries of our large
rivers, and by so doing deposit their eggs in the
purest waters at the commencement of the spawning
season before frosts and spring floods can impair the

THE SCALES OF SALMON 107

redds or destroy the eggs in the early and delicate
stages. If these high tributaries, where in many
cases the finest spawning grounds are to be
found, are not occupied by the spring fish, it some-
times also happens they are not occupied by the
quick running summer fish which follow, and at the
same time it is well known that the available lower
spawning grounds are not infrequently occupied by
successive runs of late fish, so that to some extent
one set of spawners undo the work of fish which
have preceded them by turning up gravel already
protecting eggs.

The kelts of the spring fish by getting away down
stream at an early date have also a better chance
of reaching the sea before the time at which salmon
disease makes its worst appearance. Such kelts
should be respected and carefully handled by any
into whose hands they may fall. Like the bread
cast upon the waters, they will return after many
days.

The observations of Mr. Hume Patterson * on the
subject of salmon disease show that it is possible for
an apparently healthy fish to be infected by the
real cause of salmon disease the bacillus salmonis
pestis since, if the fish has been previously diseased
in fresh water, a sojourn in the sea, although it kills
saprolegnia ferax, does not kill the bacillus. His
observations also show, however, that the entrance
of the bacillus to the tissues of the salmon is most
readily gained when fish are injured and sickly.
The organism itself has been found in river water,

* " The Cause of Salmon Disease," Fishery Board for Scotland, 1903.

108 THE LIFE OF THE SALMON

and more plentifully in certain places and under
certain conditions than others. It is natural, there-
fore, to suppose that the kelts which remain longest
in a river after spawning are, in their sickly and
often wounded state, more likely to become infected
than those fish which quickly leave fresh water for
the sea. The fungus spreads quickly when fish
begin to die, since saprolegnia flourishes much better
on dead than on living tissue. It is noticeable that
diseased fish are very commonly of large size, and
that salmon disease is most conspicuous in large
rivers. It is in large rivers that large kelts remain
longest after spawning, and there is considerable
evidence to show that these late kelts are fish which
ascended fresh water in summer and autumn. Hence
I would offer the observation that loss of fish from
salmon disease is less likely to injure the stock of a
river when the river maintains a good run of spring
fish. In the immense rivers of British Columbia,
Alaska, and California the kelts of salmon which have
ascended many hundreds of miles are reported to die
without exception ; only those fish which do not
ascend to the upper waters seem to be able to return
to the sea. The great runs of fish in those Pacific
coast rivers do not occur till late summer, but the
expenditure of energy necessary for so great an
ascent must produce a much greater drain upon the
systems of the fish than is found in any upper water
fish in our country, and in all probability exceeds
the lost of nutriment sustained by our spring fish
which remain long in fresh water. But for the
excessive fishing of salmon traps in the Pacific coast

THE SCALES OF SALMON 109

rivers, it would appear, however, that the extraordi-
nary loss of kelts can be disregarded in the upkeep
of stock.

With us, it seems to me, the value of a kelt is
not sufficiently recognised. Because a kelt is an
unclean fish and very often a nuisance to the angler,
the disposition frequently seems to be to treat the
fish with contempt, to drag the hooks out of his
mouth and to throw " the dirty brute " back any-
how. I have seen a ghillie lift a kelt as a sea fisher-
man commonly lifts a dead cod-fish, by sticking the
thumb and middle finger into the two eye sockets.
What chance of survival, one may ask, has a sickly
kelt with bleeding gills, or one subjected to treat-
ment as described ? Bacillus salmonis pestis is
waiting for him, and saprolegnia has every chance
to flourish. The fish has been reproducing his species
in the river, and is willing to do so again if he is
only allowed to go to the sea and become once more
a silvery clean salmon. If he is carefully handled,
and not only treated with respect but decorated with
a silver medal bearing a distinguishing number which
corresponds with his weight and length at time of
marking, he is, in my opinion, of much more value
than a clean-run fish which is knocked on the head
and put in the boiling pot.

CHAPTER V

THE FEEDING OF THE SALMON AND GROWTH
OF THE GENITALIA

Absence of food in salmon taken in fresh water Feeding
habits of fry Food of salmon in the sea Nutriment in
tissues of estuary as against upper water fish Transference
of nutriment from muscles to reproductive organs

THE scales and the signs of growth and of spawning,
as well as other points in the life of the salmon, are
intimately connected with the feeding and non-feed-
ing periods of the fish.

It is recognised on all hands that the stomach of
the salmon when taken in fresh water is invariably,
or all but invariably, empty. A further examination
also shows that the intestine is empty, and shows
no indication of nourishment having been taken for
some time previously ; and that the gall bladder is
collapsed. Hoek * in the lower Rhine examined
2000 salmon, and found food remains in only seven,
which were taken in the months of March and April
Meischer Ruesch t at Basle, 500 miles up the river,
examined 2162 salmon in four years in connection

* Rapport over Statistische en biologische onderzockingen ingesteld
met behulp van in Nederland gevangen Zalmen.

j- Statistische und biologische Beitriige. Zur Kenntniss von Leben
des Rheinlachses in Susswasser. (Metzger und Wittig, Leipzig, 1880.)

THE FEEDING OF SALMON 111

with his important physiological investigations, and
found food in only two male kelts. The food was
believed to be, in the one case, the remains of a
roach, and in the other the scales of a cyprinoid
fish. In our country and also on the Atlantic
coast of America, hundreds of salmon have been from
time to time examined, and always with the same
result. It is needless to dwell on this or to give
references in detail, for several writers have already
done so in recent publications. For details of the
examination of the intestinal tract of such salmon,
reference may be made to the report by Dr. Noel
Paton and others, published by the Fishery Board
for Scotland,* and to the writings of Dr. Kingston
Barton. In the investigation of the stomach and
intestine of salmon from fresh water, made by one
of Dr. Noel Paton's collaborators, a mistake was no
doubt made owing to post-mortem changes having
affected the tissues examined. From the patho-
logical condition, thus unhappily accepted as normal,
the inference was drawn that salmon were unable,
when in fresh water, to absorb nourishment if any
food were taken. Mr. W. Earl Hodgson in his book
on " Salmon Fishing " says that the Fishery Board
are committed to this desquamative catarrh theory.
This is a little hard on the Fishery Board, and,
moreover, is not quite accurate, since, after the
error had been pointed out by Dr. Kingston Barton,
sufficient acknowledgment was surely made in ask-
ing the critic to continue the investigation and to

* Report of Investigations on the Life History of the Salmon.
1898.

112 THE LIFE OF THE SALMON

contribute a paper to the Fishery Board's Reports,
which paper was published in 1902.

The main point is, however, that food is not taken
in any regular manner in fresh water, and that the
fish is in no sense nourished. In support of this ic
may be added that the marking and recapture of
clean run fish during their stay in fresh water has
invariably shown a loss of weight. I think, more-
over, that it is right here to add that those who
in our country corroborated the results of Hoek and
Meischer in this matter expressly stated, when they
demonstrated the non-feeding of fish in fresh water,
that much depended on what is meant by the word
" feeding." The " taking " of fly, minnow, or worm
was freely admitted, and it was also admitted that
worms and other objects were occasionally swal-
lowed.* The interpretation of "feeding" (Report,
p. 170) was the " digestion, absorption, and utilisa-
tion of material by the body." Some reference to
this seems necessary, for so many appear, by their
criticisms, to have supposed that, by declaring salmon
did not feed in fresh water, Hoek, Meischer, Noel
Paton, Kingston Barton, and others meant that the
taking of the angler's lure had no connection with
feeding. The Jock Scott or Silver Doctor may
certainly be like nothing in the heavens above or the
earth beneath, but the minnow or gudgeon is after
all a fish, even though it smacks of formalin. Such
lures are found by experience to attract the salmon,

* In 1885 a small trout was taken from the stomach of a grilse
caught in the river Thurso, and sent by Mr. Robertson to the Field
office.

THE FEEDING OF SALMON

113

and the impulse to take them is in all likelihood the
same impulse which enables the salmon to nourish
himself at other times. In this sense the fish may
be said to feed, while at the same time there is
nothing unnatural in allowing that the wobbling of
an apparently half-dead dace or sprat over a salmon's
head may incite the poor fish's rage, or that the ex-
quisite colouring of what is called a fly may produce
a flash of keen emotion, as has been said by some.

When the digestive tract of the salmon taken at
the mouth of the river is examined, it is natural to
suppose, since the fish is more recently from the sea
than is the upper water fish, that more trace of
feeding will be found. Grey and Tosh, in 1894 and
1895, examined 1694 salmon in the Tweed, 1442 of
the fish being taken at the mouth of the river in the
nets of the Berwick Salmon Fisheries Company.*
Of those Berwick fish 128, or 9 per cent., contained
food. The following table shows in a condensed
form the times at which the fish with food were
taken :

Month.

Number

examined.

Number
with Food.

Per Cent,
with Food.

February

7

1

14-3

March .

46

20

43-4

April .

133

53

36-9

May

215

36

167

June .

236

31

13-1

July .

283

5

17

August

210

8

3-8

September

106

2

2-0

It is noticeable at once that the greatest number

* Fourteenth Annual Report, Fishery Board for Scotland
Part II. Note 2.

H

114 THE LIFE OF THE SALMON

with food is found in the fish taken during the
months of April, May, and June. A careful ex-
amination was made as to the nature of the food,
and a detailed table is given on pages 77 to 80 of
the report referred to. Herring remains figure most
largely, while other fish reported are sand-eels,
whiting, and haddock. In a considerable number
were found crustacean remains, in a few fish marine
worms, while amongst curious oddments and these
are interesting when we recollect the nature of the
salmon fly we have a caterpillar, four feathers, a
leaf of a beech tree, moss, blades of grass, and
spikelets of sedge. The staple food, however, seems
to be the herring, which, amongst all fish in our
seas, has been shown to be at once the most nourish-
ing and the most easily digested. Concerning the
small Crustacea (mostly amphipods), we may probably
be not far wrong in assuming that they were ingested ;
in other words, that the herring swallowed the am-
phipods, and the salmon swallowed the herring.

It is interesting to notice that from the commence-
ment of the salmon's life the feeding habit waxes
and wanes with the seasons. Every angler knows
how persistently parr will keep rising to the fly, how
greedy and troublesome they are in the spring and
early summer. When fry are reared in ponds and
hand fed they show the same peculiarities which seem
to mark the "taking" proclivities of the adults.
Some time since interesting notes, taken by the
keeper of the rearing ponds at the mouth of the
Spey, to which I have already referred, were supplied
to me. During the first year in the life of the fry

THE FEEDING OF SALMON 115

food is taken freely through the summer, but when
the first frosts of autumn set in feeding becomes
intermittent. In December many days will pass
without any food being taken, and in January and
February the amount of food consumed is very slight.
If open weather comes with spring the small parr
will recommence feeding, but frosts or unsteady con-
ditions will still check them. By the month of April
the little fish are feeding steadily and vigorously,
and simultaneously their growth becomes more rapid.
All through summer this steady feeding goes on,
and it is a beautiful sight to see the shoals of active
little fish flashing to the surface as the food is thrown
in. Their appetites are those of lusty youngsters,
and food never seems amiss at any time.

In July, if the temperature of the water rises to
about 70 F., they become less keen, and if a spell
of hot weather continues feeding will cease. In
thundery weather they will not rise from the bottom
of the pond.

With the advent of winter a distinct lessening of
the desire for food again manifests itself, although
during this second winter entire cessation of feeding
for several days does not seem * to occur so fre-
quently as amongst the fry.

By the month of March the fish are two years old,
and feeding is again freely maintained. As the
silvery dress is assumed the smolts become very keen
and active, and feed at any time. Under natural
conditions they should now descend towards the
sea.

Even when a plentiful supply of food is obtainable,

116 THE LIFE OF THE SALMON

therefore, the young salmon in fresh water does not
incline to feed freely in winter ; and when the smolt
stage is passed and the fish has reached the great
feeding place the sea this peculiarity, which is
shared with other fishes, is still obvious. In the
formation and growth of the fish's scales these
periods of feeding and non-feeding are recorded, as
we have seen ; while, with regard to the length of
time during each summer in which the salmon
continues to feed, it is found, both by observations as
to food remains and by observations on the state of
nutriment in the tissues of estuary fish, that feeding
continues without cessation till September.

An interesting letter bearing upon the subject of
the food of the salmon was read by Dr. Dunlop, in
his evidence before Lord Elgin's Salmon Fisheries
Commission. It was written by Mr. Rae, one time
superintendent of the Duke of Richmond and
Gordon's fisheries, and is as follows : " When salmon
were boiled and kitted I had great facility for ob-
serving the contents of their stomachs, and I have
invariably found that in all places the character of
their food, as a whole, has been the same. At
Lochinver in 1847 ; Carlo way, in Lewes, in 1848,
1849, 1850, and 1851 ; at Loch Ewe and Loch
Broom in 1852, 1853, 1854, 1855, 1856, I had
occasion to open many thousands, and found that
the principal contents were herrings. As many as
half a dozen in one fish, more or less decomposed, in
many cases reduced to pulp ; in others as many as six
or eight herring bones decomposing. I also found
sand-eels in many cases. , . ." Day mentions seeing

THE FEEDING OF SALMON 1 17

sand-eels and herrings in the stomachs, and reports
that Jardine states that salmon are often taken on
the coast of Sutherland, on haddock lines baited
with sand-eels. Other observers have made similar
reports. Even in fish shops where sea-caught
salmon are sold, after being cut across, one may
sometimes notice that herring had been the food
last taken.

There is, then, a great contrast between the con-
dition of the stomach and intestines of fish caught
in the sea and the condition observed in fish caught
in rivers. In the former case the stomach is com-
monly found full of herring or sand-eel food, the
intestines being correspondingly filled with fsecal
matter, while in the latter both stomach and
intestines are empty. We have, therefore, well-
defined periods in the life of the fish, and long
intervals during which no food is taken.

One of the main objects of the investigations
undertaken by Dr. Noel Paton and his colleagues
was to determine whether salmon in fresh water
require food, and in the event of corroboration of
the results obtained by the Continental physiologists,
to study the changes which must necessarily take
place owing to the growth of the genitalia during a
period of starvation. In the female salmon the
growth of the reproductive organs is very great.
It has been ascertained that in April and May the
ovaries constitute about 1 "2 per cent, of the weight
of the fish, while at the spawning season they repre-
sent as much as 2 3 '3 per cent. In the male fish,
the actual increase is not in so great a proportion

118

THE LIFE OF THE SALMON

the contrast between spring fish and autumn fish is
also well marked.

In dealing with fish from different rivers, it was
necessary to calculate a standard fish length with
which to make comparisons. This having been
arrived at, not only weights of fish but weights of
muscle and genital ia could be dealt with on one
basis. In estimating the nutrition of fish in estu-
aries as compared with those from upper waters,
Drs. Noel Paton and Dunlop received material from
the Spey, the Helmsdale, and the Dee. The weight
of the muscle was worked out in every case, and the
following table showing the average weight of muscle
for fish of standard length gives the ultimate
results :

May and June

Estuary.

Upper Water

Difference.

grms
6326

grms.
5839

grms.

487

July and August .

6901

5887

1014

Oct. and Nov. . .

6055

4116

1939

In the upper water fish a loss in weight of muscle
works out to nearly 28 per cent.

In contrast to this, the increase in weight of the
ovaries of female fish is most marked. In the Tweed,
Grey and Tosh, treating this subject in percentages
to the weight of the fish, had already shown a change
from 075 per cent, in March to 17 per cent, in
November. The average weight of ovaries per fish
of standard length is given by Noel Paton and
Dunlop in this form :

THE FEEDING OF SALMON

119

Estuary.

Limits of
Variation.

Upper
Water.

Limits of
Variation.

May and June . .

121

3-175

263

104-415

July and August . .

284

109-479

510

353-691

Oct. and Nov.

1439

1310-1930

2230

1851-2888

" This gives an increase from May to November of
from 1 to 11 '9 in the sea, or 1318 grms. per fish of
standard length, and from 1 to 8 '5 in the river, or
1967 grms. per fish of standard length. It will thus
be seen that although in the upper waters a greater
amount of material per fish of standard length is laid
on by the ovaries, their rate of growth, considering
the weight with which they start in May, is quite as
great in the sea as in the river."

For the growth and development of those ovaries
a very large amount of fat is necessary, since the
constitution of the ovaries is very largely of fat.

Now the average amount of fats per fish of stan-
dard length is given for both muscle and ovaries as
follows (Report, p. 97) :

Ovaries.

Muscle.

Estuary.

Upper Water.

Estuary

Upper Water.

May and June

12

29

768

448

July and August .

18

46

770

476

Oct. and Nov.

145

204

426

159

Here we see very plainly how with the decrease
of muscle fat the ovary fat increases. Taking the

120 THE LIFE OF THE SALMON

whole season into account, the fat per fish of standard
length in the ovaries of upper water fish is about
double that in estuary fish. In the early months,
May to August, the increase is comparatively small,
but in October and November the increase is four and
a half times.

From the results obtained the deduction is " that
the fat which the salmon has stored in its muscle
when it leaves the sea is not only amply sufficient to
yield all the fat required for the fats of the growing
ovary, but also abundantly sufficient to yield energy
for an enormous amount of muscular work." In
point of fact, the amount used apparently in energy
is very great.

In dealing with male fish, it was found that the
storage of muscle fat is as great in those entering
our estuaries as in the case of females, and that this
fat is used up to as great an extent, although less
goes to the male organs than to the ovaries. The
presumption is that the amount required for energy
is greater.

Without entering upon further details which would
weary the reader, it may be said that the fat in pro-
cess of being utilised by the ovaries is to some extent
changed in character, being largely combined with
phosphorus, and that the other essential element
in development of the genitalia, the proteids (which
include the familiar " curd " of the fresh run fish),
are found to be stored in large quantities in the
estuary fish, and to be reduced in the muscle and
transferred to the ovaries in upper water fish.

Such then, in brief, is the chemistry of this matter.

THE FEEDING OF SALMON 121

But in order to increase the scientific value of the
demonstration a thorough microscopic investigation
into the physiological conditions of the muscles
and their contained fats was undertaken by Mr.
Mahalanobis. In the highly charged condition of
the estuary fish the fat cells were found not only
between the muscle fibres, but within the fibres
themselves. In the upper water fish the gradual
disappearance of the fat was most clearly established,
thus supporting in every way the deductions drawn
from the wider chemical study.

From these investigations it appears, therefore,
that the feeding which the salmon does in the sea is
much more than enough to enable the fish to grow
and develop in the manner usual with animals of a
steady feeding habit. The feeding is so excessive
that the health and energy of the fish is more than
supplied, and as a result the excess of nutriment is
stored in the tissues. Reference here has alone been
made to the muscular system, but it has to be under-
stood that the abdominal organs and their connecting
membranes are likewise loaded with fat under ordi-
nary conditions. After the cessation of feeding and
the entrance of fresh water this source of nourish-
ment is drawn upon for the supply of the reproductive
organs and for the supply of energy necessary to
the fish in ascending rivers and performing the
reproductive functions.

That this supply for energy is ample will be
readily understood by the fisherman who has had the
keen joy of testing from the end of his rod the sort
of fight which a well-nourished salmon can put up.

122 THE LIFE OF THE SALMON

That salmon do not invariably enter fresh water
in an excessively over-nourished state I have already
noted. A season occurs now and again during which
fish in comparatively poor condition leave the sea
for our rivers. This in all probability means that
the nature of the food taken in the sea has been
poor in quality, or, less probably, scarce. Further,
the common occurrence of so-called bull-trout enter-
ing our rivers from the sea, pale in flesh, often of
great size, and having both the gill maggots common
to the kelt and the sea lice of the fresh run fish
upon them, suggests the belief that amongst fish
which adopt the short migration there are those
which do not adopt the generous feeding habits of
the silvery well-conditioned migrant.

Such considerations as these seem to me to prompt
a certain reserve in accepting the conclusions of Dr.
Noel Paton that it is the state of nutrition alone
which is the factor determining migration towards
the river. We have dwelt at considerable length,
in the chapter on the results of salmon marking, on
the habit of long and of short periods of sojourn in the
sea. Since a salmon in any year can clearly become
fully nourished during the time of a short sojourn in
the sea, why should it stay an extra year in salt
water ? The state of its nutrition does not supply
a sufficient reason. The corollary to the view
advanced by Dr. Noel Paton is expressed by him
thus : " When the salmon has accumulated the
necessary supply of material it tends to return to its
original habitat." That the salmon is a native of
fresh water, in the sense of being born there, I of

THE FEEDING OF SALMON 123

course freely admit ; but, as suggested in the chapter
on grilse, it is possible to regard the first seaward
migration as a movement to the salmon's original
habitat, i.e., the habitat from which our salmon as we
now know it, along with its near allies, took its origin
in remote times. The shad is a sea fish which visits
our rivers to spawn. The sparling or smelt must be
similarly regarded, and is more nearly related to the
salmon than the shad. Attempts at producing
fertile salmon without allowing them to visit the
sea easily fail, and when successful, result in
degenerate creatures which do not support the idea
that they represent the prototypes of fish from
which the evolution of our salmon has sprung. The
vigorous growth and development in the sea, on the
other hand, and the fact that the salmon spends
most of its time in this element, along with the
arguments already cited in the chapter on grilse,
from the systematic study of the salmon family, in-
cline me to the view that the sea was the original
habitat of Salmo solar, as it undoubtedly was of
most of the salmon family.

If this is so, there is nothing forced in the belief
that the long migration fish, continuing in the sea
as it must do long after its tissues have become fully
nourished, and as the study of the scales certainly
indicates, is simply living in its natural habitat. At
the same time, I do not consider that the nisus
generativus prompts the spring fish to enter fresh
water. The late running fish which has already
developed its genitalia in the sea may be said to..
visit the river for the purpose of spawning, but the

iyr.

124 THE LIFE OF THE SALMON

species Salmo solar I regard in general terms as a
sea fish which freely enters fresh water after it has
received the nourishment and stimulus the sea
alone affords, which frequently, owing to the highly
charged state of its nutrition, spends many months
in fresh water, and which spawns before again re-
turning to the sea.

CHAPTER VI

THE SALMON AND WATER TEMPERATURE

Early and late rivers Rivers made late by man Tempera-
ture of the sea round Scotland Temperatures of early and
late Scottish rivers Temperature of water no indication of
seasonal character Effects of temperature on fish ascending
rivers Early runs of fish in upper waters, Garry and Orchy
Effects of cold water upon ascending fish The entrance
of tributaries from main river Rapid ascent of summer fish
Sea lice may remain attached for four or five days

THE theories which have been advanced to explain
the diverse seasonal runs of salmon into fresh water
why in some rivers there should each year be
early runs of fish, and in others only late runs
have, with I think one exception, been based upon
considerations of water temperature. The exception
is the ingenious but unsupported theory of the late
Frank Buckland, that the seasonal character of a
salmon river depends on the length of the river in
proportion to the square mileage of its drainage area.
At the outset, in the consideration of this question,
I would point out that we are at the present day by
no means certain that the real seasonal character of
some rivers can be properly interpreted by the exist-
ing conditions. Certain rivers are described as late
rivers, and have no doubt been found to be late

126 THE LIFE OF THE SALMON

rivers for a very long period ; but it is quite possible,
and I would venture to say in many cases very
probable, that the cause of the late character has
nothing whatever to do with the physical conditions
of the rivers, but is purely the result of the influence
of man in his treatment of the fisheries, and that as
the treatment has continued on the same lines within
the memory of man, so no one has ever known or
heard of early runs of fish in those rivers. One might
instance the case of the river Leven, which flows
out of Loch Lomond into the Clyde at Dumbarton.
This river has for a very long time been considered
a late river, and there are old records of Dumbarton
fisheries which point to this conclusion. A thriving
dyeing and bleaching industry has long been estab-
lished, gross pollution has resulted, and net fishing
at the mouth of the river has been steadily carried
on. But of late years an active and I may add a
philanthropic angling association, which allows all
anglers to participate in the sport which the loch
affords, has done something to improve the general
conditions and the stock of fish, with the result that
early fish are appearing. In other cases where
serious over-fishing has been allowed to continue
rivers have naturally acquired a late character. With
over-fishing, or other causes of reduction of breeding
stock, the first class of fish to disappear is the spring
run. We are therefore not safe in presuming that
in all cases when we do not find a spring run the river
is a late one. We must satisfy ourselves first that the
river has every chance of bearing a full stock of fish.
In Scotland and concerning questions of tern-

THE SALMON AND WATER TEMPERATURE 127

perature in relation to salmon fisheries I write
exclusively of Scotland we may say broadly we
do not find a regular run of spring fish in small
rivers, even although these rivers are well stocked,
and pure, and in no way over-fished. In rivers of
moderate size we, for the most part, find only
occasional runs of rather late spring fish. But in
the larger and more important rivers, Tay, Dee,
Ness, Beauly, Brora, Helmsdale, Thurso, Naver, we
have regular and steady runs of spring fish. In these
rivers also we have, besides considerable volume of
water, comparative purity in the Highland rivers
named we may say complete purity and an apparent
absence of over-fishing. In a river like the Tweed
there is too much netting to allow of the entrance of
many springers, or of their passage upwards to the
safer head waters. Other rivers might be mentioned
which like the Tweed should, in my opinion, have
many more spring fish; but to mention individual cases
would be to depart from the question of temperature.
The fact that the first-class rivers above named
flow either to the east or north has suggested the
theory that the cause of the early entrance of salmon
lies in the fact thatj the sea on this side of the
country the North Sea is a colder sea than the
Atlantic Ocean on the west coast, and that the
rivers, flowing, as many of them do, through long
stretches of not very high land, are comparatively
warm, being free from the great admixture of snow
water which in spring passes off from the high West
Highland hills, and that the salmon therefore prefer
to leave this cold North Sea at an earlier time than

128 THE LIFE OF THE SALMON

do the salmon in the warmer Atlantic. The theory
sounds plausible, but in accepting it one has to pre-
sume that the salmon prefers warm water to cold.
The more crude idea that early rivers are those
which flow out of large lochs is an offshoot of the
same theory, and may be dismissed at once by merely
noticing how many early rivers do not come out of
large lochs, and that practically all the late rivers
of the West Highlands do. A third idea which has
been developed at considerable length, more especi-
ally in magazine articles, is that the ascent from sea
to river takes place when the temperatures of the
salt and fresh water approximate. We have there-
fore the physiologist stating that the condition of
nutrition determines the " return " from the sea,
and the physicist stating that the influence of
temperature alone determines the same movement.
I have already referred to the question of nutrition,
and have shown that in my view it is not a complete
explanation of the habit of ascent from sea water.
In examining the theories as to temperature it is re-
markable how much of theory and how little of fact
is to be found. Only in arguments as to the ap-
proximation of sea and river temperatures do we find
a genuine attempt to deduce from thermometric
readings. When such readings are taken in sufficient
numbers to show the real conditions which usually
obtain, it becomes at once evident that the idea that
salmon are drawn, as it were, from a cold sea by and
to a warm river must be given up. The facts show
that at the time when early fish are running all our
rivers are much colder than the sea on either coast.

THE SALMON AND WATER TEMPERATURE 129

A most exhaustive series of sea temperature records
are available, and, as shown at some length in the
nineteenth Annual Report on the Salmon Fisheries
of Scotland, the monthly means for January, Feb-
ruary, March, and April range on the east coast from
39 to 44 F., the lowest monthly mean of the whole
year being in March, while on the west coast it is evi-
dent, on the high authority of the late Dr. Buchan,
that " from October to March the west is at least
2 above the east, and in January it amounts to 3*9.
It is an event happening only once in a number of
years for the temperature of the sea in the west to
fall below 40, but in the east this happens every
year." * In the west the three coldest months of
the year are in order February, March, and January.
The statement that the North Sea is, as compared
to the ocean water on the west coast, a cold sea in
the early months of the year, when spring fish run,
is therefore supported by facts. It is when we ex-
amine river temperatures during those early months
which the theorisers referred to omitted to do
that we find an entire absence of any warm water to
draw fish as supposed.

A series of maximum and minimum morning and
evening readings taken from instruments kept con-
stantly immersed for four years 1880-1884 in-
clusive in the Helmsdale and Brora, early rivers
of east Sutherland, were kindly placed at my dis-
posal by the Scottish Meteorological Society. From
these we find t that the Helmsdale and Brora waters

* Jour. Scot. Meteor. Soc., vol. i., N.S., 1863-66, p. 263.
f Nineteenth Annual Report, Fishery Board for Scotland, II.
p. 70.

I

130 THE LIFE OF THE SALMON

are in January and February from 5 to 10 colder
than the sea. In March and April the conditions
become more unsettled, and sudden and rather
violent fluctuations are common (melting snow, frost,
rain floods). A variation of 8 in the twenty-four
hours is not uncommon. On an average, however,
the river water is 4 colder than the sea. The
Thurso is another well-known early river. Tempera-
tures for the complete year 1886 were taken at
three points, at the mouth, half-way up, and at the
head, together with sea temperatures for the same
period.* 1 In January and February the river mean
is 6 to 9 colder than the sea, and not till April do
the river and sea temperatures equalise. Readings
from the Dee show the same result, as also do
readings from the Tay and Shin, although from
these rivers the series of temperatures is not so long
as from the Helmsdale and Brora. A series of
valuable readings for ten years taken in the Ugie
bear out, however, a precisely similar result, f Lest,
however, it should be imagined that, with the
greater fluctuation in river than in sea, periods
occur, even in the early months of January and
February, when river temperature equalises with
sea temperature, it may be said that, the daily
readings show that the river water is constantly
colder than the sea. In those east coast rivers,
therefore, the early running fish leave a cold sea for
a considerably colder river in their migration to
fresh water.

* Jour. Scot. Meteor. Soc., III. No. iv., 1886.

f Nineteenth Annual Report, Fishery Board for Scotland, II.
p. 72.

THE SALMON AND WATER TEMPERATURE 131

With regard to the west coast rivers, which were
supposed to be colder in the early months of the year,
I may cite first the Laxford in west Sutherland. I
have complete readings from this river, taken in
1880, and am therefore able to compare it with the
Helmsdale, Brora, and Shin in east Sutherland in
1880. The Laxford flows out of Loch More, is con-
sidered the finest salmon river of the North-West
Highlands, but is late. A comparison of the daily
means of those rivers shows that the Laxford is sub-
ject to rapid fluctuations, as the Helmsdale is, and in
this respect differs somewhat from the Brora and
Shin. The latter, it must be recollected, has a very
large reservoir at its head in the shape of Loch Shin,
while the actual readings for the former were taken
at a constriction of Loch Brora. When, however,
curves are plotted representing the daily means of
each river,* it is noticeable that they do not separate
from one another in any marked way, and that the
inclination of the Laxford curve, the western late
river curve, is very similar to that of the Brora
curve. In other words, the temperature conditions
ascertained to exist in the Laxford are very similar
indeed to the temperature conditions of the three
early east coast rivers with which it is compared.

The great advantage of being able to compare
late and early rivers on opposite coasts of the same
country in the same year is further enhanced by the
fact that the Helmsdale, Brora, and Laxford are
first-class natural rivers, each bearing a full stock of

* Twentieth Annual Report, Fishery Board for Scotland, II.
p. 12.

132

fish, each unspoilt by man. In a search for a really
warmer river, I have been fortunate in obtaining a
valuable series of temperature readings from the
Nith, which flows into the Solway. These were
taken for me by Rev. Mr. Andson, of Dumfries, in
1902, and I am able to compare them with Tay
readings for the same year. In the early months of
the year the Nith is distinctly warmer than the Tay,
and the Nith is a late river, but unhappily it cannot
be said to be a river in which the natural supply of
salmon is unimpaired. The salmon fisheries of the
Nith district have suffered much at the hands of
man. In this same year, however, 1902, tempera-
tures were taken in the river Ness, which is a well-
stocked early river, supplied from one of the largest
and deepest lochs in Scotland, and the Ness tempera-
ture shows a marked difference from that of the Tay.
Curves have been plotted for each river,* from which
it is at once noticeable that while the Ness shows a
comparatively flat curve never falling below 42, the
Tay curve at Perth shows acute inclination with a
minimum at freezing point (32).

We see therefore, from a comparison of the ascer-
tained thermal conditions of those rivers, that east
coast early rivers do not show any marked difference
from late west coast rivers when the selection is made
in Sutherland alone ; and second, that east coast
early rivers in different parts of the country, subject
to different physical conditions, show widely different
temperatures with minimi as much as 10 apart. It

* Twenty-first Report, Fishery Board for Scotland, II. pp. 76
and 82.

THE SALMON AND WATER TEMPERATURE 133

has therefore seemed to me that the natural deduc-
tion may be said broadly to be that temperature has
nothing to do with the seasonable character of a
river in Scotland, in so far as the actual entrance of
salmon from the sea is concerned. The primary
causal factor in the existence or absence of an early
run of fish is not the prevailing condition of tempera-
tare. Once a fish has entered the mouth of a river
fluctuations of temperature exercise a distinct in-
fluence, as we shall presently see, but in the making
of an early or of a late river temperature does not
appear to play a part. Large and deep lochs
naturally exercise their influence on the temperatures
of rivers which flow out of them such lochs have
relatively high mean temperatures ; but apart from
such conditions, the temperature of a river follows
close, so to speak, upon the heels of the air tempera-
ture of the surrounding district ; and if a warm
river were the chief cause of early runs of fish the
small streams of the western isles should of all others
be early streams, and the rivers in the south-west of
Scotland should also be all early rivers. As observed
at the commencement of this chapter, however,
small streams are never early streams. A certain
volume of water and a good stock of fish are
essentials for the early run. Given these two condi-
tions, the time at which early fish will enter fresh
water may still be subject to some variation in
different districts. On the east coast and in the
Pentland Firth salmon may draw to the shore rather
earlier than they do on the west coast, the habit in
the sea where circumstances determining the growth

134 THE LIFE OF THE SALMON

and movements of marine creatures, including salmon,
and the forms upon which they feed, may be in some
slight degree different, but given these two con-
ditions an early run will be present. The main
circumstances of life in a small country like Scotland
are, however, practically the same.

We may now regard the influence of temperature
upon the salmon when the mouth of the river has
been entered. In this connection there are two
aspects of the question which may be considered,
first the ascent of the main river, and second the
entrance of tributaries. Apart from the test of
temperature, it will be recollected that at the close
of the last chapter the results of marking clean run
fish were dealt with, and evidence produced to show
that such fish may remain two or three months in the
rapid lower waters of a river such as the Spey with-
out making material ascent. From the Tay similar
evidence was gathered, but the pause in the ascent
appeared to be less prolonged. In now extending the
range of observation, and considering the application
of the temperature test to this question, we are able
to see that there are other rivers in which this
pausing habit does not apparently obtain. The habit
of the early fish seems to vary considerably in different
districts. We have already noted that Tay fish ascend
to Loch Tay and remain there till May, when they
continue their ascent by entering the Dochart. As a
contrast to this we have early Ness fish passing quickly
through the short river Ness into the loch of the
same name (which is the deepest but one in Scotland),
but, unlike the Tay fish, passing on out of Loch Ness

THE SALMON AND WATER TEMPERATURE 135

into the river Oich, out of the Oich into Loch Oich,
and out of Loch Oich into the famous river Garry,
which forms the head waters of the Ness system.
The Garry is the spring fishing river of this district.
On the west coast we have an analogous condition
appearing in the Awe district. Early fish pass
through the river Awe into Loch Awe, and are
checked in this ascent only after they have entered
the river Orchy above the loch. The Garry and the
Orchy are rivers of considerable size, although each
is some distance from the sea. In each there is a
fall, beyond which fish do not commonly pass till
about May. Now, why do Tay fish remain in the
lower river and loch during the spring months, and
Ness and Awe fish ascend to the upper feeders of
their districts ?

In order to ascertain the thermal conditions of the
river systems referred to, I arranged for the taking
of morning and evening maximum and minimum
readings from thermometers kept constantly im-
mersed in the Ness and the Garry and in the Awe
and the Orchy. At the same time temperatures
were taken in the Tay and in one or two smaller
river systems. The Garry rod-fishing opens early
in February, and at that time each year the river is
well stocked, although in exceptionally cold springs
there maybe more fish in Loch Oich than in the Garry.
In the Awe, fish do not usually ascend till March,
and are not expected in any numbers in the Orchy
till April. It is highly probable that with a fuller
stock of fish in the district more earlier fish would
appear. The readings procured from these districts

136 . THE LIFE OF THE SALMON

are of great interest as showing that in each case
the lower river which flows from the large loch into
the sea has a markedly higher temperature than the
river above the loch. It is noticeable also, when the
charts which were prepared * are studied, that the
curves representing the mean weekly temperature
of the Ness and Awe (the lower rivers) are more
uniform in line or less quickly fluctuating than are
the curves for the upper waters. The Ness curve
during January, February, and March shows a
wonderfully flat inclination. The temperatures were
taken November 1901 till end of May 1902. I will
refer first to the Garry and Ness readings. From
the middle of December to the end of March the
weekly mean of the Garry never exceeds 40 F., and
in the middle of February, when the river was very
low after four weeks of continuous frost, the reading
reaches 35. In April and May there is a steady
rise of temperature. The curve for the Ness runs,
however, at a distinctly higher level. In Decem ber
January, February, and March there is constantly
from 2 to 8 difference in the readings. By the
middle of April the temperatures of the two rivers
have approximated by the rise in temperature found
in the Garry, the upper and early river. On referring
now to the curves representing the Orchy and Awe
temperatures, we notice first a much greater fluctua-
tion in both curves, but the same general result.
The Orchy is colder by 2 to 9 than the Awe, and
by the middle of April the curves approximate by

* Twenty-first Annual Report, Fishery Board for Scotland, II.
p. 76.

THE SALMON AND WATER TEMPERATURE 137

the rise in temperature of the lower line, which
represents the Orchy or upper water.

Now, from a constant record of temperatures taken
in the Tay at Grandtully by Mr. Johnston between
1895 and 1903, coupled with statements as to the
results of spring angling, it has been possible to pro-
duce a table showing the maximum and minimum
readings for each of the early months of the years
referred to, and to compare the seasons in respect
of the ascent of spring fish.* The year 1896 was a
good spring fishing year, when "fish were early and
ran quickly." Only once, in January, did the
temperature reach as low as 36, the lowest readings
in February, March, and April being respectively
39, 40, and 43. The highest readings for the four
months were respectively 41, 44, 43, and 51.
The average of the maximum readings is 44*7, and
of the minimum readings 39'5. Poor fishing years
when compared to this show at least 3 lower tem-
perature in both maximum and minimum averages.
In other words, with the river temperature com-
monly above 40 in the first four months of the year
salmon are running up freely and quickly, and the
stock of spring fish is largely dependent upon favour-
able conditions for ascent in January before the
opening of the netting season. This is exactly the
condition we find in the Ness and Awe, and we are
not surprised to find that the early fish of those
rivers run quickly through the few miles of water
(each river is only about six miles long) into the
large lochs, where the temperature of the water is

* LOG, cit. t p. 78.

138 THE LIFE OF THE SALMON

not greatly different, except possibly at the extreme
upper ends. When, however, the fish essay to
ascend farther, they are met by distinctly colder
water, and their progress is at once checked, till, the
particular time when the winter conditions of those
upper rivers having passed, they find themselves
once more willing to travel and overcome obstacles
in their course. The rivers Garry and Orchy are
the early rivers of their districts, therefore, because
all the lower waters are of such a relatively high
temperature as induces rapid ascent, while the Garry
and Orchy are of a temperature which in Scotland
produces a check upon early ascent. Added to this
we must not forget that small falls exist in each
case. Spring fish are no jumpers till summer con-
ditions have come about.

Apply this interpretation to the Tay and its upper
water, the Dochart, and we see that while the main
river, with its course of some 50 miles, receives
various large tributaries, and has a general tempera-
ture in the first four months of the year which at
times induces fish to run and at times checks their
running, thus naturally securing to this fine river a
distribution of spring fish, the Dochart which has
a fall upon it a short distance above Loch Tay is
exactly analogous to the waters of the Garry or
Orchy above their respective falls. Fish therefore,
if we argue by analogy, should, as in the Garry and
Orchy, ascend the Dochart fall when the wintry
conditions have passed. This is, as a matter of fact,
exactly what happens, for the Loch Tay fish begin
to go up the Dochart in May.

THE SALMON AND WATER TEMPERATURE 139

Reference in this connection may be made to one
other river, the Helmsdale, which, since it differs in
physical features from the rivers already referred to,
may stand as a further example of how this concep-
tion of salmon ascent as affected by temperature
works out in Scotland. Here there is no large loch
in the course of the main ascent. Two streams come
from two lochs at practically the extreme head waters,
and unite to form the main river, which has then a
course of about 20 miles to the sea. About half-way
down this main river, however, there is a fall. What
we find here is that the early fish enter freely from
the sea and slowly make their way upwards, but do
not ascend the fall. All the spring fishing is between
the fall and the sea. When, however, the wintry con-
ditions of water have passed, the ascent of the fall
is freely made, as in the upper waters previously
referred to, so that summer angling is carried on in
the entire length of the river. Falls such as we have
been referring to are therefore not serious obstruc-
tions to fish when the water is comparatively warm,
but are total obstructions when the water is cold.
Such a statement, I must again remind the reader,
applies to Scotland ; it would, I believe, be entirely
erroneous in a country like Norway, where all sal-
mon, although late, ascend obstacles in the cold
water which comes from melted snow. Salmon passes
on such Scottish falls, to be of any use, must be of
extremely easy gradient, so that fish may readily
swim them, and at the same time they should con-
tain a large body of water.

We have now to consider the part water tempera-

140 THE LIFE OF THE SALMON

ture seems to play in the entrance of tributaries by
salmon which have ascended a main river. This
subject has been chiefly studied in the Tay, and no
river in Scotland seems to offer better facilities,
since the tributaries of the Tay are numerous, and
are in several instances rivers of considerable size?
which present varying physical conditions. Read-
ings were taken in the main river at Perth which
is at the mouth, but at the top of tide reach and
at Grandtully about forty miles up stream. Above
Grandtully the river Lyon enters, and in this
readings were taken. Below Grandtully the river
Tummel enters, carrying with it the water of the
Perthshire Garry, and in both of these were readings
also taken. Curves were plotted for each series of
readings. Of the two from the main river it is
noticeable that the curve from the mouth of the
river is lower than that for the upper river. This
is no doubt caused by the entrance of the large
tributaries below Grandtully, the Tummel and Garry,
and the Isla, but especially perhaps the Almond, the
waters of which are found to be colder than the
waters of the other Tay tributaries. The Lyon
alone shows a curve of temperatures very similar to
the Tay. All other tributaries for which readings
are available are found to be colder. Loch Tay 9
from which the main river springs, is 14|- miles long,
and has a mean depth of 199 ft. and a maximum
depth of 508 ft. The water which flows from this
large loch is naturally much more uniform in tem-
perature than is the water of rivers draining high
surrounding land. A chart of curves showing

THE SALMON AND WATER TEMPERATURE 141

weekly means for the Tay at Perth and at Grand-
tully, and for the Lyon and Garry, is given in
Part II. of the Twenty-first Fishery Board Report,
and it seems unnecessary to enter here upon a
detailed account of the readings from each locality.

In general terms the results show, as compared to
the main river, the Grandtully curve of which is
comparatively high, that in descending order of
temperature we have first the Lyon, then the
Tummel, and last the Garry. It is therefore of some
interest to realise that early fish in ascending the
Tay are found on the opening of the angling season,
January 15, not only in the loch, but in the river
Lyon, which, as we have seen, has a temperature
closely approximated to that of the upper Tay and
loch. Next, a limited number of fish are found to
be already entering the mouth of the Tummel and
pausing in the deep pools of Easthaugh Water there,
but that as the temperature of the colder Garry
rises so as to approximate to that of the Tummel,
fish more freely ascend the Tummel and then enter
the Garry. This last emigration does not usually
take place till April, although a few fish may be
found in the deep pools of the Pass of Killiecrankie
in March. Moreover, in unusually warm springs,
when all the tributaries are, as it were, advanced in
temperature, the times at which fish ascend are,
while unaltered in order of river, earlier in point of
season ; and conversely in cold seasons the stock of
early fish is delayed in their ascent of the tributaries.
If therefore in early winter much snow lies on the
hills drained, say, by the Lyon, so that that river

142 THE LIFE OF THE SALMON

has a temperature a few degrees lower than the Tay
water, a good stock of spring fish may not be ex-
pected in Glen Lyon at the opening of the season ;
but, other things being equal, it is found by actual
thermometric readings that the order in which fish
enter the Tay tributaries, in cases where free access
is at all times possible, is precisely the order of
descending temperature. To state the case in more
general terms, it may be said that the tributaries
referred to are not entered by early fish till their
temperatures have become approximated to the
temperature of the main river near the influx of
the tributaries.

It might further be noticed that the tendency of
the early salmon in making his steady ascent of such
a river as the Tay seems to be to pass by tributaries
which enter the lower reaches, and especially to pass
by tributaries of small size. I know of no instance
in Scotland of early fish entering tributaries near to
the mouth of the main river, as autumn fish un-
doubtedly do. In the Spey, for instance, the tribu-
taries of Fiddoch and Avon are frequently crowded
with late running spawners, but they never contain
spring fish. In the Tay the tributaries Almond and
Isla, both large enough to hold spring fish, are passed
by, but there may be other reasons, since the former,
as we have seen, is a cold river, and moreover is too
much obstructed to make the ascent of spring fish
possible, while the latter, for which I have no tem-
perature records, is provided with a poor mouth.

In other cases which might be cited there is
some uncertainty in determining whether a river

THE SALMON AND WATER TEMPERATURE 143

should be considered a tributary or a separate river
entering a common estuary. This is the case with
the Earn, which enters the tidal waters of the Tay,
and but for obstructions in the Earn I believe that
spring fish would ascend more freely than they do.
In the case, however, of the Lochy on the west
coast we have another instance worth referring to.
The river Lochy formerly flowed naturally from Loch
Lochy into the sea near Banavie ; when, however,
the Caledonian Canal was constructed the course of
the Lochy was to some extent utilised, and the river
was given a new and artificially cut channel from
the loch into the Spean at Muccomer. The Spean,
coming as it does from the very high land to the
north and east of Ben Nevis, is a cold river, while
the Lochy water, like the Ness at the northern
end of the great chain of lakes which form the
Caledonian Canal, is warm. Early fish ascend the
Lochy to Muccomer, and there pause. They are,
unfortunately, obstructed by a fall with a very
inadequate fish pass from farther ascending the
Lochy, so they remain in the large Muccomer pool
till the Spean water has lost its forbidding tempera-
ture, when they ascend that river. Such fish as do
manage to ascend the Lochy do not do so before
the ascent of the Spean is well begun. The reason
for this has already been explained. We see, there-
fore, that while early as compared with late rivers
appear to present no definite distinction of tempera-
ture, the habits of ascent in all rivers are very
considerably influenced by thermal conditions. All
our Scottish rivers are colder than the surrounding

144 THE LIFE OF THE SALMON

seas in the winter months, but by the month of
April have risen in temperature so that the readings
in fresh water and the sea become equalised. Sub-
sequently, through summer, our rivers attain a
temperature considerably above that of the sea, but
by the month of September their temperature has
again fallen so that the conditions of sea and river
again equalise although now at a higher tempera-
ture than in .April. From September onwards the
winter conditions again become slowly established.

During the warm conditions of summer, fish
ascending from the sea run up much more rapidly
than when the river temperature is low. Even in
May, fish are not infrequently taken well up our
rivers, i.e., thirty to even forty miles up, and still
bearing sea lice upon them, and with other unmis-
takable signs of having only very recently left the
sea. By this time the spring fish are making their
way to the head waters, where, as the season ad-
vances and their reproductive organs develop, they
become distributed in the manner so valuable for
the river's future stock. The summer fish having
rapidly moved up from the lower waters, the late
running fish, which enter from the sea with their
genitalia already well advanced the male fish being
at the end of the season not infrequently highly
coloured, as they come from the sea have, as it
were, those lower reaches left to them. In this way,
therefore, the waters of a well-stocked pure river are
fully taken advantage of for spawning, and are able
to return their best percentage of fry from all classes
of fish.

THE SALMON AND WATER TEMPERATURE 145

In the summer of 1906 an experiment was made
at my request in the laboratory of the Fishery
Board at Bay of Nigg, near Aberdeen, to test the
length of time sea lice will remain attached to those
summer fish which so rapidly ascend in the compara-
tively warm conditions referred to. Two grilse were
obtained from a bag-net in the Bay of Nigg, and
placed in sea water in a tank of the hatchery. The
temperature of the water was 52'9(11 < 6C.). Each
fish had attached to it a number of sea lice. The
density of the water was then reduced by allowing
fresh water to enter. This operation was regulated
so as to represent approximately in time the period
of a single flood tide. The fish showed considerable
distress at first, from which it is natural to suppose
that the transference to brackish water was too
rapid, and that in all probability an interval of some
twenty-four hours or so as had originally been
intended would have been more natural.

The sea lice may also have been adversely affected
by the rapid transference in eight hours to pure
fresh water, but in spite of this some of the sea lice
remained attached for six days. Dating from the
time when the water was quite fresh, the sea lice
remained on one fish for four days and on the other
for five days. The fresh water temperature was 55 F.

The constant action of the current of a river
under natural conditions of a fish's ascent may cause
the parasites to drop off some what sooner, but the ex-
periment shows that a fish taken in upper waters with
one or two sea lice attached may have occupied at
least three or four days in its passage from the tide.

APPENDIX

A RECORD OF SALMON-MARKING OBSERVATIONS

IN SCOTLAND PREVIOUS TO THE OPERATIONS

OF THE FISHERY BOARD IN 1896.

MARKING salmon for the purpose of gaining information as to
their migratory movements and their increase in weight has been
practised for a considerable time. Izaak Walton in his " Compleat
Angler" (1653) alludes to experiments of this kind in which
ribbons were tied to the tails of young salmon, and by which it
was ascertained that salmon return to the same locality after
visiting the sea. "Writers of a later date do not always specify
the method of marking, although they state results which would
be of great interest were the proofs of their infallibility absolute.
A considerable amount of reliable information has, however, been
obtained from time to time in Scotland exclusively, I think, on
migration and increase of weight and I have attempted to collate
this information in the form of a list with tables. I omit from
the list references to smolt marking, as well as any records which
are without such particulars as alone can render them of value.

PREVIOUS SCOTTISH SALMON MARKING.

1823. In March, Mackenzie of Ardross marked a grilse kelt of

3 Ib. The mark was of brass wire.

In twelve months the fish was recaptured as a salmon
of 7 Ib., showing an addition of 3 Ib. in weight.

1824. Mackenzie of Ardross marked a salmon kelt. The mark

was wire round the tail.

In 1825 he reports it as having been recaptured
" double the size."

148 APPENDIX

1829. In February, Fraser (" On the Salmon ") reports marking a

number of grilse kelts, estimated as of 5 to 6 Ib.
Marking was done by cutting off the adipose fin.

In nine months he reports the capture of one as a
salmon of 13 Ib., showing an addition of 8 Ib. in weight.

This return must be received with caution.

1830. In February, Mr. Fraser (ibid.*) reports the marking of

grilse by fastening wire round the tail. Grilse pre-
sumably 5 to 6 Ib.

In six months one was recaptured as a salmon of
14 Ib., two witnesses being cited who saw the fish and
wire.

1834. Shaw at Drumlanrig marked 524 sea trout in the herling

stage. Marking was done by fin cutting. One fish is
reported to have been recaptured in the three successive
years in a cruive in the Nith. The markings were :

1834, adipose fin removed.

1835, third of dorsal fin removed.

1836, half of anal fin removed.

1837, fish was killed,

being afterwards exhibited at the Royal Society of Edin-
burgh, vide Trans. R.S.E., vol. xv. part iii. p. 369 ;
weight 6 Ib.

Of the 524 marked in 1834, 68 were reported recap-
tured in 1835, averaging 2| Ib. in weight.

1835. The 68 recaptured sea trout above referred to were re-

marked by fin cutting.

One in 20 are reported as recaptured in 1836, averaging
4 Ib. in weight.

Also 120 sea trout were marked by Shaw in this year,
60 being marked with copper wire on the dorsal fin,
and 60 with copper wire round the right maxillary bone.

Of the latter group five recaptures were made in the
summer of 1836, the average increase of weight being
stated as 1^ Ib.

1841. Young of Invershin marked grilse and recaptured salmon
and as under (Trans. Royal Society, Edinburgh, read

1842. January 9, 1843). The manner of marking is un-
fortunately not specified,

APPENDIX

149

LIST OF GRILSES MARKED IN THE RIVER SHIN, AND RECAPTURED
AS SALMON ON THEIR SECOND ASCENT FROM THE SEA IN
THAT RIVER.

MARKED.

RECAPTURED

WEIGHTS WHEN

Marked.

Recaptured.

February 18, 1841

June 23, 1841

4 lb.*

9 lb.

February 18, 1841

June 25, 1841

4

11

February 18, 1841

June 25, 1841

4

,

9

February 18, 1841

June 25, 1841

4

10

February 18, 1841

July 27, 1841

4

13 .,

February 18, 1841

July 28, 1841

4

10

March 4, 1841 .

July 1, 1841

4

12

March 4, 1841 .

July 1, 1841

4

14

March 4, 1841 .

July 27, 1841

4

12

January 29, 1842

July 4, 1842

4

8

January 29, 1842

July 14, 1842

4

9

January 29, 1842

July 14, 1842

4

8

March 8, 1842 .

July 23, 1842

4

9

January 29, 1842

July 29, 1842

4

11 ,

March 8, 1842 .

August 4, 1842

4

10

January 29, 1842

August 11, 1842

4

12

1845. Lord Glenlyon, afterwards Duke of Atboll, marked on

March 31 a kipper fish of 10 lb. It was marked with

a zinc label, and is reported by Ramsbottom to have

been recaptured in less than six weeks weighing 21 J lb

(? W. L. C.)
1847. March 27, the Duke of Atholl marked a 15 lb. kelt in

the Tay. It was recaptured on February 20, 1848.

as a clean fish of 20 lb.

March 28, marked a kelt of 11 lb. It was recaptured

on July 24, presumably the same year, as a clean fish of

18 lb.

March 3, marked a kelt of 14 lb. It was recaptured

on September 7, presumably the same year, as a clean

fish of 21 lb.
1851. The Duke of Atholl marked, on February 26, a kelt of

10 lb. It was recaptured, a clean fish of 12 lb., on

February 17, 1852.

* Grilse of weights other than 4 lb. were rejected.

150

APPENDIX

1854. The Duke of Atholl marked an 11 Ib. kelt on March 21.
It was recaptured as a clean fish of 17 Ib. near Montrose
on August 24, 1854.

1851 During those years a large number of fish were marked by
to the Experimental Committee of the Tweed Commis-

1872. sioners. The table which follows (pp. 152 to 156) is
taken from the Tweed Salmon Reports, 1866. It is
given in full, although it contains reference to a few
marked smolts. I have added a column as to the
nature of the mark used, so far as this can be determined
from the Report.

1857. On November 12, Mr. Buist, Stormontfield, river Tay

while netting for ova, marked 89 salmon and 78 grilse
with copper wire (where on the fish is not stated).

On September 6, 1858, one was caught in Largo Bay,
13^ Ib. in weight. No particulars of each fish seem to
have been taken.

1858. March 31, Young of Invershin reports that a grilse marked

on that date was recaptured as a salmon of 8 Ib. on
August 2 in the same year.

1859. The Duke of Atholl marked fish on the river Tay. A

numbered copper disc, an inch in diameter, or gutta-
percha ticket fastened by copper wire round the tail was
used. The following fish were marked as kelts and re-
captured as clean fish the same year :

HARKED.

DATE.

WEIGHT.

LENGTH.

DATE OP

RECAPTURE.

WEIGHT.

No. 21

Feb. 14

10 Ib.

36 in.

Aug. 18

17 Ib.

76

Mar. 2

11 J Ib.

35J in.

Aug. 18

17 Ib.

95

Mar. 29

12J Ib.

37 in.

Aug. 12

19 Ib.

1860. Through thei kindness of the present Duke of Atholl I am
able to add particulars of another fish, the mark taken
from which I have seen. I add also the lengths of
the three fish above referred to.

APPENDIX J51

No. 133, marked on Februai-y 12, 1860, a kelt, 8 Ib.
caught at The Rock, Dunkeld, 34 inches long by 12
inches girth. This fish was taken again in the river
Tay, but the mark and the particulars at time of recap-
ture were not sent to Blair Castle.

1861. Also, on February 25, a kelt of 13 Ib. was marked. It
was recaptured as a clean fish of 26 Ib. on August 11,
1862.

1863. On March 30, at Dunkeld, one of the Duke of Atholl's
marks was attached to a kelt 5 Ib. in weight and 28
inches long. It was recaptured at Kossie, in Montrose
Basin, on July 17 of same year a lanky male " bull
trout "of 7lb.

152

APPENDIX

1

i

S

o be

o

-

M

o

> a

\

M

3

c3

03 V

s

s

S

|

d js

1

ff

c

1

I

fer

SB

1

s

1 "

NATURE

be

5

bn

i

_a

*-

LNumbered

iS ^

sg

3S

*^

s

Wire in tail

Numbered
culum.

Stiver wire

'

Probably w

rfi

H

'

H

X3

M

^

1

r*.

t

S

1 rf

I

5

CO

- *.

.='

.0

, 1

Q

H

HC i-

af o

J5 J

9 t~

"

^

o

w

CO

^

CO

f

J

!

CD

CD"

ID

* 3"

M

73 1

: -c

1

g

^

c

1

cS

B

|

O C

5 O

O

H

H

O

W

CO

O

JS

WHERI

Wilford

Ethermout

Finchey

Newmills

Yarmouth

Shields
Eyemouth

Wilford

Finchey

Ethermout

Miwlowlm

Goswick

d

j.

00

o

ti

1
a

g

* CD

^

ja

tf

49

X

a

0}

S

H

K

3

d

I

"3

|

c!

|

" 1

3

i

N)

to

' I

GO

w

t

^*

e4

IO

<D

M

10

IO

00

00

"

CO

00

"

00

"

" "

M

7!

3 *

'

S

I

p

H

M

I

g

1

^5

5

"s
a

K

1

I

H

CO

CO O

O

1

i

s

"

1

g

"3 .

1

ft

5

j
.
W C

"o

a "3
1 1

T =

D i

5 a

<a

r

M

1

olt of one

1 -trout B
Floors

trout 8m

1

CO

1

CO

O

a c

co E

t -3

H QQ

P

'f
EH

s

a

CO

S

S

CO

"3

CO

i

d

'

6

9

^

E OF MARE

fe
t

r I

c

'

05
CM

|

s .

bt

a

CO

1

1

1

s s

H

,4"

M

*"

MS*

IO

Q

CO

"

"

*" "

"* **

APPENDIX

153

S3

k)

1

"^

1

Q

s

d

A

2

b

gd

V /

be

S

S " S

&

"3

d

B

a

o * .

3 d

M

k

Q

I

'

- on

- 1

a

"3 d

O 1

N

S

a

6*5 =

M

D

liver wire

r

i

utta-percl

=

alvanised
part of "

opper or
in " nose
ilver wire

fc

CO

G5

O CO

^

H

l

2

S

,0

94

.

^a

g

o

S =

' ' J

CO

CO

.0

01
O

"^

X MH. ^ Zw

p

j

"

3
8

"3
%

00

d"
o

3

be
d

~ CT. CO CO

50 d"

*T o

H

PQ

a
c5

=3

M

S

1

1

I 1 *

Q

WHERE.

Carr Bock .

Hallowell .

Hallowstell

Abbotsford

Carbarn

d
K

Aberdeen

Hornclifle .
North Esk

= 5

PQ 1 -a o 5
2 pd

CO CO M

'on, and Cove. W

a

Q

cs

00
94

_

CO

I

W
1

10

-g

1-1

to

00

!c to

f-I 'to

5

9

o P

|

1

1

P.

1

"3

a 1

1 a a ^

.0
5**

H

otT

* ^j-

M

a

'

ee

CO CO

ao

'

S

00

rH 2

|

g

1 '

-

i

Q

1

O

B
OQ

8

QQ

*

W

O

+

c3

o

c^**

M

g

Q

p

(M

m

a

m

M

Bull-tront Smolt .

Salmon Smolt

Bull-trout Smolt ,

H

8

}

Grilse Kelt 2 Ib.,
New Water

o ^

11

Three " Blacktails,'

" Blacktail," 8 tol:

i)

"d

"3

M .

2 5 ''
H <!

S 3

d

M

M

S

3 OF MAR

1C

1

K

r

d

March 31

J

a

Autumn

October

be
o

1

CO

B

1^"

06"

cT

cT

i

00

K

S

>o

00

K

B

R

o

CO "

154

APPENDIX

3

HE OF M.,

B

IE

1

a

1

SS, '-

K

s

3

2

K

1

^

~

a

B

SJ

3

a

a

h

g
fi

|

E

. . . 1

S

I

c

2

,c

c

f ' '

i

HN

*

1

2

*

;

pi

H-f.
CO

HN

CS ,2 T3

.Q

CN

I

1 &

Q

*

1

o "2

^

f

1 W

_

S 60

bn

2 *bo

M

Blacktai

Clean T

o a

'If

O g

H

1

u

1 1

H f>

isi

H ^

g

O

u

.

E

w . S

.

.

,

.

WHER

1

Wiliord

Needle Ey<
Meadowha

1

03 'S O

(H W H! H

Tillmouth

Carr Rock

JJ

|j

'

R

Q

S fe

IH

a

00 -w ^

' *

fi '

3

"

<M

**.!

0>

*

f i

,p

3

"3

= s

s ' ^ S

"3

I i

f *

") i-s

K

CO

CO

10

CO

<o

00

CO

"

" CO

a '

fl

s

0*

A

8>

K

a

O
g

3

KI

.

.

s

KINDS

"Orange-fin "o
Smolt .

"Blacktail"

"3 ~
W

a
2 '
A

tf

"3
M

s s J 7

c '

"Blacktail"

=

" Silver-white

= S

at

"* H

K
M

E OF MAR

fi

Autumn

ft

CO

s r K s

October

-

Autumn

be

1 1

s; oci

&

CM

1C

CO

CO

P

00

"

00

. -

"

.

H
525

APPENDIX

155

_ 13

8 fl o
6 08 S

od ^

S

*

M FH

IN

C4

1- O>

O> Ud

i-l eo

e,

CO

"^ .2 P.

Q

ll

|S

1

-H

O

rl

CO l~

^

-

| 3

1

S a

i

a

s r

a

E E

s s

E ?

3

row

1

1C

1

HM

H4

H*| i

1 1 MH

1

Q *

1-1

a

M

8 .
a _g

N

O

d

O

1

~

1-1

|

<u

a

N

O

1 t

8 1 5

s 2 1 8

1

S ,?

IN

CO

-ff

g

iH D rH

s sr

S p

rt

2

Jj

Q

P

Sag

^'"M

00

1

2 5

1

00

S 5

s s"

"o" 1

i i ra

HS

4B

.fi

g-

1

.

-

CO N
H O

- !

O i-l

I 1

| o

N

i i

I

-o

<-? F-fS*

IN

> "

^

,Q ^5

* K

P

<N

1-1

IN

"*

1

Where
Retaken.

Hallowstell.

Galashiels.

North Esk,
Montrose.
Cove.

West Ord.

Broad.

Coquet,
Warkworth.
Sandstell.

Watham.
Hallowstell.

1 1

Watham.

English New
Water.
Ethermoutb.

Scotch New
Water.

bo some doubt.

SH ^

w

IN ,

"O

MS

10

^

e us

rl to

t-'

a

IN

IN i-l

1-1

1-1

CO "^

C* r-t

1-t

e * ei

P.

S p,
d u

FH "3 1-1

t 00

P. >

rt

gp

f; a

b oli

2 2

r= B 2 OJl

tf

P M

fc

CO <

^

3 j

^

fe ^5 ^

^0

H

Species of Fish
when
iRecaptured.

Whitling.

Salmon.

Bull-trout,
do.

Yellow Trout.

1
i

do.
Whitling.

Blacktail.
Bull-trout.

do.
Whitling.

j tlD
O *

Blacktail.

articular case se

3 a

^

^

CO 00

OS

as

-X CO

10

H*

P,

2

CO

rl

ex IH

C4 t*

"'

H

.

I

,0

N

O

5

s

O

OTHI

5?

c*

^* ^

CO

WHi iHN

^J* HN H.

t r-l

p

IN

"""

Where
Marked.

1

CO

|

Scotch New
Water.

Start.

Yardford.

Ethermouth.

South Bells.
Ethermouth.

Scotch New
Water.

Start.

Scotch New
Water,
do.

Start.
Low Bells.
Ethermouth.

Scotch New
Water.

til

o> a

I-l

1^1

CO

O

o -o

1

-d

4

o

O

o oo a *c
1-1 a

i

1

^j

j

Species of Fi

Wbitling.

Salmon.

Blacktail.
Bull-trout.*

a
2

H
|

"3

Bull-trout.

Do.
Blacktail.

Bull-trout.
Whitling.

|*|

C a |s
CD 8 o

o i

Blacktail.

156

APPENDIX

*||

8?
&*

P

S

-

S

CO

<N

1

-

-

o

CO

IN

S 1

1

> "55

> O

m

*

fc.

OO

oo

10

C-.

00

1

G

oo a

CO

S *

g

6+3
a to

=

r

I

1

r

s

S

s

a

S '

'- K

fi" S

He

CO

i i-tn

CO

H

HM

H

(M

CO

CO

HO

i o ^

IO

S ^

Q

D

1

e

5 ^f
S ^

O
>o

l-

o

o

a, N
14,

o

O

O 4

j-

09 c

O

1

O

o

gS =
HS H

Q

.5
-t*

_

HOI

HO

H5

w*

HC<

00

co

KS

r-t

TJ1

CO

^

00

t-

,5

Jj

O

e

I

.0

t.

N

O

A

e*

o

C

J

o

S e

^

H*

.0

HCT

oo

7?"

.0

ow

00

CO

c*

.=

c?

CN

00

ll

1

C

a

eS
O

S '<

H

Fallin, near
Stirling-on-

Yardford.

Sandstell.

Hollywell.

Shoreside.

Broad.

i

-a

H

English New
Water.

'j.

V

a
J

Watham.

sll

H H

Finchey.

H g

O

co

o

00

co'

.

00

tf

ci """*

rH

^*

^

CO

*~*

CO

~v

c^ " H

94

C4

SS

e! o

3 1"

to

a

!

c

03

a

1-5

9

"5

K

C

! H

>

si

"I

>

A

.3 ^

fe ^

s|!

to

_a

d

1

1

.*

1

d

d

'3

|

d

d

'

to
3 d

I

1*1

3
p

"3

a

I

S

a

a

a

i

n

^s
P

a

i
s

3 "*

O

CO

r

HCT

Hn

H5

HO

ft-S

*

o

CO

CO

>*

1Q

t-

Td

r-l Oi

Jl

A
to

J

o

o

N

O

o

o

o

N
O

N

O

N

O

b
TI

N N

!

o

.^

CO

CO

00

CO

CO

CO

O*

CM

o

*t

KHi

*

1-1

"^

rH

*"*

<N

.=

1-1

CO

k

^3

! 3

|

Where
Marked

02

1^

1

CO

'

O

o

s

1

s

Horn well

6

a

Eithermou

1

CO

Horn well

fl

s

CO

Low Bell
Eithermou

* .
fe

S!

8^

*r*

s^* a

S

O

O

o

a

o

a

t
o

o
a

1

d

o

3

o

o

d d

a -a

oo ja

a

rt o

!z;

I

ei

M

"R

s

C3

o>

1

1

O
P

d
Q

Q

o
Q

o
Q

Q

Q

<5

P

O

5

1

Q Q

O

1

?

-

00

02

H
525

APPENDIX 157

it will be noticed from a study of the table that
" blacktails " are recaptured both as whitling and as
bull-trout, but that fish marked as representatives of
either of these latter stages are in each case recaptured
as whitling and bull-trout respectively. That is to say,
so far as the table shows, a whitling does not become a
bull-trout, even, as in the case of the first fish men-
tioned on the list, after an interval of about a year and
ten months. The fish referred to seems, on recapture,
to have weighed 2 lb., and may therefore be more
correctly described as a sea trout. We may perhaps
fairly infer from this result of marking " blacktails "
that the young stages of the bull-trout and of the
whitling were indistinguishable to the observer who
conducted the marking, or are practically indis-
tinguishable.

Another fish worthy of special remark is the last on
the list, a grilse which on being captured after an
interval of six months is still described as a grilse, and
has gained 4^ lb. in weight. From the date of mark-
ing we may infer that the fish was a kelt, and that we
are dealing with a minimum fresh water weight as
contrasted with a weight after feeding.

1873. The Tweed Committee conducted further marking observa-
tions. The results, published as a special report in
1875, deal almost exclusively with "blacktails" re-
captured as blacktails and whitling. The only excep-
tions are two grilse recaptured as grilse kelts before
leaving the river.

INDEX

ABERDEEN fishery, grilse and salmon

takes, 34, 38
Archer, W. B., emigration of small

salmon 49

salmon marking in Norway, 57
Awe and Orchy temperatures, 135

BERBINGTON, A. D., over-netting re-
sults, 106

Bornholm, line fishing for salmon, 35
Boulenger, marine origin of salmon, 5
British Columbia, death of kelts, 108
Brora, marked fish recaptures, 71

grilse migrations, 55
Buchan, Dr. , temperatures of Scottish

seas, 129
Bull trout of Tay and other rivers, 87

CAITHNESS, takes of salmon and

grilse, 38
Curd in salmon, 120

DAHL, K., capture of smolts and

grilse off Norway, 24, 26
age of Norwegian smolts, 11
Deveron grilse, migrations of, 54
Dickson, J., divided migrations of

salmon, 59

Divided migrations of salmon, 59
Dunlop and Noel Faton, nutrition of
salmon, 118

EARLY and late rivers, 125
Early rivers not necessarily of high
temperature, 130

FEEDING, absence of in fresh water,

43, 112
in sea, 43

Fishing Board for Scotland, salmon
marking experiments of, 57

GENITALIA, growth of, 117

Gilling, 41

Grey and Tosh, feeding of salmon in

Tweed, 113

Grilse, marked smolts recaptured
as, 30

not entering rivers, 87

early, 34

on coast, 34

and small spring fish, 39

in sea-water ponds, 9

in poor condition in rivers, 42

scales of, 40

recaptured as salmon, 70

genitalia of, 53

HATCHING, influence of temperature,

13

Helmsdale, migrations from, 55
Herring, chief food of salmon, 114
Hodgson, W. Earl, non-feeding in

fresh water, 111

Hoek, grilse and salmon in Rhine, 51
absence of food in Rhine fish, 42
Holt, spring fish in Ireland, 90
Hume Patterson, on salmon disease,

107

IRISH salmon marking, 63

JOHNSTON, H. W., scales of grilse, 40
scales of Tay salmon, 96

KELTS, abuse of, 109

of spring fish, 106

weight in Scotland and Ireland, 6S

descent of, 85

Kingston Barton, digestive tract of
salmon, 111

MALE* fish, their wanderings, 79
Marking of smolts, 3, 29
of salmon results, 56,

160

INDEX

Meischer Ruesch, absence of food in

Rhine fish, 110
Migrations, peculiar, 79
Migratory movements on coast, 50, 82

NESS and Garry temperatures, 135

Netting, importance of moderation in

rivers, 76

of salmon in sea, 77
and reduction of stock, 17

Noel Paton, fresh water origin of

salmon, 44
satiety of fish, 42
and Dunlop, nutrition, 118

Nutrition of salmon leaving sea, 43

PETER of the pools," 2
Parr, descent of on Pacific coast, 11
descent of at Stormont field, 2
scales of Norwegian, 11
of upper tidal water, varied size,

21

and smolts, feeding habits, 114
Patterson, J. Hume, salmon disease,
107

REARING of smolts, 4

Rutter, on parr of Pacific coast, 11

SALMON, small spring, 39

probable marine origin, 44

entering other rivers, 78

return to their own rivers, 78

marks now in use, 58

marking, 56

small spring, scales of, 40

disease, 107

record weight of Scottish, 106

fishing by set lines in Baltic, 35

Salmonidae, majority marine, 44

Saprolegnia ferax, 107

Scales, development of in fry, 95
of salmon as records of life, 93
of marked smolts and spring fish,

100

of marked smolts and grilse, 99
lines of growth, 104

Sea lice, life of in fish in rivers, 145
trout smolts, 5

Smolts, uniformity in nature, 18
prevented from migrating, 7
destruction of at river mouth, 21
congregate in upper tidal waters,

19

two years in fresh water, 7, 17
eggs from ripe females, 8
or young grilse iu sea, 25
autumn migrations of, 22
confined in sea ponds, 9
food of in sea, 24
habits in sea, 27
marking in Tay, 29
scales, lines of growth, 98

Spawning seasons of salmon, 12
habits of salmon, 14
annual and biennial, 60
mark on scales, 97
infrequency of, 102, 105

Spring fish, ascending habits of, 89
spawning habits, 106

Stormontfield experiments, 2

Sutherlandshire rivers, temperature
of, 129

TABLES of salmon marking results,

63

Tay, records of marked salmon, 70
Temperature of rivers, 123
Thurso, temperatures In spring, 130
Tosh and Grey, Tweed grilse, 42

feeding of Tweed salmon, 113
Tributaries, entrance of by salmon,

139

Trout, feeding in estuaries, 5
in New Zealand, 6
estuarial varieties, 6
Tweed Commissioners, marking ex-
periments, 57

WATER temperatures, 125

Weight, increase between spawning

periods, 84
Willis Bund, proportion of grilse in

Tweed, 51

^Printed by BALLANTYNE & Co. LIMITED
Tavistock Street, London

Telegrams : 4 1 and 43 Maddox Street,

1 Scholarly, London.' Bond Street, London, W.,

March, 1908.

Mr. Edward Arnold's

List of New Books.

NEW NOVEL.

By the author of ' The Jungle.'

THE METROPOLIS.

By UPTON SINCLAIR.
Crown Svo. 6s.

This is Mr. Upton Sinclair's first new novel since ' The Jungle,'
wherein he startled the world with his lurid account of the iniquities
perpetrated in the American canned meat trade.

In ' The Metropolis ' he brings a scathing indictment against the
vulgar element in New York Society, whose exotic and vicious lives
are exposed with all that unrestrained and brilliant relentlessness of
which Mr. Sinclair is a master.

It will be remembered that a few months ago a report went round
to the effect that Mr. Sinclair had managed to get a situation as
butler in the house of a prominent millionaire, where he was studying
the occupants with the intention of describing them in a book. The
report was unfounded, and was denied at the lime ; but it is quite
clear from this book that he must have had opportunities of no
ordinary character to have enabled him to handle the subject with
such telling and convincing effect.

LONDON : EDWARD ARNOLD, 41 & 43 MADDOX STREET, W.

2 Mr. Edward Arnold's List of New Books

NEW NOVEL.

By the author of ' The House of Shadows. 1

THE WAYS OF REBELLION.

By REGINALD FARRER.
Crown 8vo. 6s.

Mr. Farrer has gained for himself a reputation as a writer of novels
with fresh and original themes. In the present instance certain
phases of modern life are hit off with great skill and incisiveness.
The vacillating relations of a young married lady with a man who
has divested himself of his wealth in order to devote his life to service
among the poor, are handled in a manner extremely modern. Mr.
Farrer knows his world, and provides much crisp dialogue and clean-
cut phraseology.

FROM THEIR POINT OF VIEW.

By M. LOANE,

AUTHOR OF ' THE NEXT STREET BUT ONE,' ' THE QUEEN'S POOR,' ETC.

Crown 8vo. 6s.

Miss Loane is a district nurse ; she has lived among the poor and
for the poor ; she knows the society of the poor from the inside, yet
she comes in from the outside, consequently she sees closely enough
to descry details accurately.

This new book, full of real knowledge, common sense, and robust
humour, is urgently recommended to the attentive perusal of all men
and women who are interested in the problem of life among the
poor. The following are among topics discussed in it : The
Manufacture of the Tramp, The Cost of Food, Mental and Moral
Characteristics, Why the Poor prefer Town Life, What is Charity ?
Miss Loane has a great gift for telling anecdotes, and employs it
with effect in the present instance.

NEW AND CHEAPER EDITION.

THE NEXT STREET BUT ONE.

By M. LOANE.
Crown 8vo. 33. 6d.

Mr. Edward Arnold's List of New Books 3

GEORGE RIDDING, SCHOOLMASTER
AND BISHOP.

3fort2*tbfro 1beao /toaster of Wtncbester, 1866-1884.
fftrst ffiisbop of Soutbwell, 1884-1004.

By his Wife, Lady LAURA RIDDING.
With Illustrations and a Plan. Demy Svo. 155. net.

Whilst appealing especially to Wykehamists, this valuable
and interesting biography claims the attention of the larger public that
enjoys reading about the careers of successful and distinguished
people.

The volume is roughly divided into three sections, the first telling
of Dr. Ridding's early days at Winchester and Oxford. After this
comes the account of his notable headmastership notable because
during it was evolved the Winchester known to most of her sons
^now living. So many buildings were added and so many improve-
ments made that Dr. Ridding, the prime mover in it all, came to be
known as ' the second founder.' The third section describes the
formation of the Diocese of Southwell, and is a remarkable revela-
tion of the successful application of exceptional administrative
ability.

PICTORIAL ART IN THE FAR EAST.

By LAURENCE BINYON.

With Coloured Frontispiece and numerotis other Illustrations beautifully
reproduced in collotype. Crown 4^0. ais. net.

This important work, which is only rendered possible by the
immense additions to our knowedge of Far Eastern art during the
last decade, brings out and establishes the high interest of Chinese
painting, hitherto practically unknown in Europe, and of the older
schools of Japan, the subsidiary schools of India, Persia and Tibet
being also glanced at. The author's aim has been to treat his
subject not merely from the technical historical side, but as a theme
of living and universal interest, with its background of Oriental
thought and civilization.

4 Mr. Edward Arnold's List of New Books

MINES AND MINERALS OF THE
BRITISH EMPIRE.

33eina a Description of tbe Ibtstorical, ipbgsical, ano 3ntmstrfal

features of tbe firincipal Centres of Mineral iproouction

in tbe JBritisb dominions begone tbe Seas.

By RALPH S. G. STOKES,

LATE MINING EDITOR, Rand Daily Mail, JOHANNESBURG, S.A.

Demy 8vo. With numerous Illustrations. 155. net.

This work is the outcome of a careful and exhaustive inspection
of the mines all over the Empire, covering a considerable period, on
the part of the author, who has visited personally the places he
describes, and who has brought much practical knowledge to bear
upon his subject. A moment's consideration enables one to realize
the great amount of romance there is attached to this prosaic-
sounding industry : the histories of some of the mines are capable
of filling volumes by themselves. On the other hand, the book is
indispensable as a work of reference for capitalists, investors, and
many others interested in the different aspects of mining.

TURKEY IN EUROPE.

By SIR CHARLES ELIOT, K.C.M.G.

('ODYSSEUS').

A New Edition, with Additional Chapters on Events jvom 1869 to the

Present Day.

Large Crown 8vo. 75. 6d. net.

Although the identity of ' Odysseus ' has for some time been an
open secret, it is satisfactory to be able at length to reveal definitely
the authorship of this important work. The additional chapters
contain a valuable review of the present position of the Turkish
question, and bring up to date a book that is already regarded as a
standard authority on its subject.

Mr. Edward Arnold's List of New Books 5

BOOKS RECENTLY PUBLISHED.
FROM THE NIGER TO THE NILE.

By BOYD ALEXANDER,

LIEUTENANT, RIFLE BRIGADE.

Two volumes. Large Medium Svo. With Illustrations and Maps.

365. net.

1 This is a great book on a great subject. The expedition of which it tells
the story was by far the most remarkable feat of recent African travel, the
greatest feat of endurance since Mr. Grogan's memorable journey from the
Cape to Cairo, and more fruitful in scientific results than any expedition since
that of Stanley. Mr. Alexander modestly disclaims literary skill, but we know
few books of travel written in purer English. 1 Spectator.

' It is a book which is worthy to be ranked with the classics of Stanley,
Speke, and Livingstone a book that deserves to, and assuredly will, take rank as
one of the most fascinating revelations of savage Africa. We have seldom read a
better book of travel, never a more deeply interesting one.' Country Life.

MEXICO OF THE TWENTIETH
CENTURY.

By PERCY F. MARTIN, F.R.G.S.,

AUTHOR OF 'THROUGH FIVE REPUBLICS OF SOUTH AMERICA.'

Two volumes. Demy Svo. With Illustrations and Map. 305. net.

ACROSS PERSIA.

By E. CRAWSHAY WILLIAMS.
Demy Svo. With Illustrations and Maps. 125. 6d. net.

RAILWAY ENTERPRISE IN CHINA.

Bn account of Its ciflin anD Development.

By PERCY HORACE KENT.
Demy Svo. With Maps. 125. 6d. net.

Mr. Edward Arnold's List of New Books

MODERN STUDIES.

By OLIVER ELTON, M.A.

ENGLISH LITERATURE IN THE UNIVERSITY

Large Crown 8vo. 75. 6d. net.

By OLIVER ELTON, M.A.,

PROFESSOR OF ENGLISH LITERATURE IN THE UNIVERSITY OF LIVERPOOL,

MEMOIRS OF MISTRAL.

Rendered into English by CONSTANCE ELISABETH MAUD,

AUTHOR OF 'WAGNER'S HEROES,' 'AN ENGLISH GIRL IN PARIS,' ETC.
Lyrics from the Provencal by ALMA STRETTELL (MRS. LAWRENCE HARRISON).

Demy 8vo. With Illustrations. 125. 6d. net.

THE
GROWTH OF MODERN NATIONS,

a IbfstovB of tbe iparticulartst fform of Society.

Translated from the French of HENRI DE TOURVILLE
by M. G. LOCH.

Demy 8vo. ias. 6d. net.

AN INTRODUCTION TO CHILD
STUDY.

By W. B. DRUMMOND, M.B., C.M., F.R.C.P.E.,

AUTHOR OF ' THE CHILD : His NATURE AND NURTURE.'

Crown 8vo. 6s. net.

THE CHILD'S MIND : ITS GROWTH
AND TRAINING.

By W. E. URWICK, M.A.
Crown 8vo. 45. 6d. net.

Mr. Edward Arnold's List of New Books 7

THE MYSTERY OF MARIA STELLA,
LADY NEWBOROUGH.

By SIR RALPH PAYNE-GALLWEY, BART.

Demy Svo. With over 20 Illustrations and a Photogravure Frontispiece.
Second Impression. 75. 6d. net.

MY ROCK-GARDEN.

By REGINALD FARRER,

GARDEN OF ASIA,' ' THE HOUSE OF SHADOWS
STREAMS,' ETC.

Large Crown Svo. With Illustrations. Second Impression. 75. 6d. net.

AUTHOR OF 'THE GARDEN OF ASIA,' 'THE HOUSE OF SHADOWS,' 'THE SUNDERED
STREAMS,' ETC.

THE LIFE OF THE SALMON.

TJfllltb "Reference more especially to tbe ffisb in Scotland.
By W. L. CALDERWOOD, F.R.S.E.,

INSPECTOR OF SALMON FISHERIES FOR SCOTLAND.

Demy Svo. With Illustrations. Second Impression. 75. 6d. net.

IN OUR TONGUES.

Some Dints to "KeaDers ot tbe Bnglisb SJtble.
By ROBERT H. KENNETT,

CANON OF ELY CATHEDRAL; REGIUS PROFESSOR OF HEBREW IN CAMBRIDGE UNIVERSITY, ETC.

Crown Svo. 35. 6d. net.

THE GOLDEN PORCH.

a asooft ot Oreefc ifairs Gales.

By W. M. L. HUTCHINSON.

Crown Svo. With Illustrations. 55.

Mr. Edward Arnold's List of New Books

A GALLERY OF PORTRAITS.

Repvoouceo from rtginal ;6tcbfngs.

By HELLEU.

With an Introduction by FREDERICK WEDMORE.
Crown Folio. 255. net.

FAMILIAR FACES.

By HARRY GRAHAM,

AUTHOR OF ' RUTHLESS RHYMES FOR HEARTLESS HOMES," ETC., ETC.

Medium Svo. With 1 6 Illustrations by GEORGE MORROW. 35. 6d. net.

OUT OF CHAOS.

B personal Storg of tbc devolution in "Russia,
By PRINCE MICHAEL TRUBETZKOI.

Crown 8vo. 6s.

RECENT FICTION.

Crown Svo. 6s. each.

HIS FIRST LEAVE.

By L. ALLEN HARKER,

ENTION OF THE DUKE,' '
SING PAUL AND FlAMMEl

Second Impression.

AUTHOR OF 'THE INTERVENTION OF THE DUKE,' 'WEE FOLK, GOOD FOLK,
' CONCERNING PAUL AND FIAMMETTA,' ETC.

THE ELECTION OF ISABEL.

By RONALD MACDONALD,

AUTHOR OF 'A HUMAN TRINITY,' 'THE SEA-MAID,' ETC.

THE DESERT VENTURE.

By FRANK SAVILE.
LONDON : EDWARD ARNOLD, 41 & 43 MADDOX STREET, W.