i

i

4

V'

}\

t

fk

\

Division of Fktm,

U. S. National I4ga»wm

5M

r

A ?

DEPARTMENT OF COMMERCE

BULLETIN

OF THE

UNITED STATES BUREAU OF FISHERIES

VOL. XXXII 1912

HUGH M. SMITH

COMMISSIONER

WASHINGTON

GOVERNMENT PRINTING OFFICE I9H

1

CONTENTS

Page.

Age at maturity of the Pacific coast saemon of the genus Oncorhynchus. By Charles

H. Gilbert. (Document 767, issued March 20, 1913) 1-22

A NEW SPECIES OP TROUT FROM LakE TahoE. By John Otterbein Snyder. (Document 768,

issued December 31, 1912) 23-28

Description of a new darter from Maryeand. By Lewis Radcliffe and William W. Welsh.

(Document 773, issued May 24, 1913) 29-32

The sense of smeee in fishes. By G. H. Parker and R. E. Sheldon. (Document 775, issued

May 3, 1913) 33-46

The fishes of the streams tributary to Monterey Bay, Caeifornia. By John Otterbein

Snyder. (Document 776, issued July 24, 1913) 47-72

Anatomy and histoeogy of the aeimentary tract of the king salmon. By Charles W.

Greene. (Document 777, issued July 25, 1913) 73-100

Notes on the naturae hosts of fresh-water mussels. By Thaddeus Surber. (Document

778, issued June 28, 1913) 101-116

Observations on fish scales. By T. D. A. Cockerell. (Document 779, issued October 25,

1913) 117-174

A NEW method for the determination of the food value of proteins, with application To Cynoscion rEgaliS. By George F. White and Adrian Thomas. (Document 784,

issued February 18, 1914) 175-182

Properties of fish and vegetable oil mixtures. By George F. White and Adrian Thomas.

(Document 785, issued February 18, 1914) 183-198

The effect of water-gas tar on oysters. By Philip H. Mitchell. (Document 786, issued

February 18, 1914) 199-206

The oxygen requirements of shellfish. By Philip H. Mitchell. (Document 787, issued

February 18, 1914) 207-222

The Anthozoa of the Woods Hole region. By Charles W. Hargitt. (Document 788,

issued April 25, 1914) 223-254

The Cephalopoda of the Hawaiian Islands. By S. Stillman Berry. (Document 789,

issued June 24, 1914) 255-362

Carcinoma of the thyroid in the Salmonoid fishes. By Harvey R. Gaylord and Millard

C. Marsh. (Document 790, issued April 22, 1914) 363-524

A LIMNOLOGICAL STUDY OF THE FiNGER LAKES OF New York. By Fdward A. Birge and

Chancey Juday. (Document 791, issued October 27, 1914) 525-610

HI

J

AGE AT MATURITY OF THE PACIFIC COAST SALMON OF THE GENUS ONCORHYNCHUS

By Charles H. Gilbert

Professor of Zoology, Stanford University

CONTENTS.

Page.

Introductory 3

Sockeye, or red salmon 5

King salmon, or Chinook ii

Silver salmon, or coho 15

Dog salmon 17

Humpback salmon 21

Summary of results 21

2

AGE AT MATURITY OF THE PACIFIC COAST SALMON OF THE

GENUS ONCORHYNCHUS.

By CHARLES H. GILBERT,

Professor of Zoology, Stanford University.

INTRODUCTORY.

During the past two summers (1910 and 1911) the writer has been engaged on behalf of the United States Bureau of Fisheries in determining the age at maturity of the five species of Pacific coast salmon, together with such other facts of their life history as can be inferred from the structure of their scales. The problem has a peculiar interest as concerns these species because of the fact, now sufficiently demonstrated, that all individuals perish when they have attained sexual maturity. They spawn but once and then die, whatever the age or size at which sexual maturity has been reached and whatever may be their physical condition at the time of spawning. This peculiarity in their life history renders the question of their age at maturity an unusually important one, both from the scientific fish cultural and the purely economic standpoints.

Many attempts have been made to solve the problem for the two species commer- cially most important the king salmon and the sockeye usually by marking the artificially reared fry by clipping one or other of their fins before they are liberated, in the hope of recognizing the adult fish on their return at maturity. Unfortunately, these important experiments have lacked adequate supervision, and have furnished somewhat discordant and uncertain results. These have been supplemented, however, by the experiments in Tomales Bay, Cal., and in New Zealand, where fry were planted in streams not frequented by the species in question and the return of the adults was noted. Also, in the case of the sockeye, we have had as additional evidence the quadrennial increase in the Fraser River, which has been very generally accepted as demonstrating a four- year cycle for this species.

All the evidence available was considered by F. M. Chamberlain, in his highly valuable “Observations on Salmon and Trout in Alaska” (Bureau of Fisheries Docu- ment No. 627, 1907), with the resulting conclusion (p. 66) :

There seems to be sufficient reason for believing four years to be the usual term of life for the sockeye and the king salmon, but experiment has pretty conclusively shown that they may mature in less time or may be retarded beyond that term.

a With the possible exception of certain male king salmon fingerlings, which mature precociously in the streams during their 6ist year, at a length of 3 to 7 inches. The fate of these has not as yet been determined.

3

4

bulletin of The bureau of fisheries

This may be accepted as an adequate statement of our belief respecting these species up to the present time. It has lacked in definiteness and in detail. No con- clusive evidence has been presented concerning the age of the markedly undersized fish or “grilse,” which are conspicuous parts of the king salmon and sockeye runs. We have no knowledge concerning the extreme age which any species may attain, nor concerning the proportions in which different ages are represented in the spawning run. Furthermore, the question of size with relation to age has been wholly undetermined, the belief being yet widely entertained that no such relation exists, size being solely dependent on richness of feeding in the sea.

There remains to be noticed a recent attempt by Prof. J. P. McMurrich to solve these problems by a consideration of scale and otolith markings. The figures of scales and otoliths presented in this paper show sufficiently that these structures present seasonal markings which when correctly interpreted will afford reliable indication of age. It is to be sincerely regretted that Prof. McMurrich’s interpretations have been based on inadequate data, and these have misled him into announcing a series of conclusions so largely erroneous as to confuse rather than clarify the questions at issue. A further discussion of these matters will appear under each of the species considered below.

The present paper contains a purely preliminary statement of certain facts in the history of our salmon which can be substantiated through a study of scale structure. This method offers certain obvious advantages over that of determining the age by marking the young, although the latter method should be employed and rigidly super- vised in corroboration of the former. But the scale method is of unlimited application. Any desired number of individuals can be investigated in connection with size and sex and other modifying factors. It thus becomes possible to analyze an entire salmon run into its age components, and each of these can then be the subject of further investigation and analysis. We can thus determine the range in size of each group and of the males and females separately for each age, and when definite events in the life history of the fish record themselves upon the scales as we believe to be the case corresponding cate- gories can be formed and the possible effects of these events on growth and period of maturity can be determined. A very wide usefulness, therefore, attaches to this method, and for its proper elaboration discriminating study of many thousand specimens must be made.

While the method is new as regards Pacific salmon, it has been experimentally tested and fully approved by the Fisheries Board for Scotland in the case of the Atlantic salmon, and is now universally accepted as furnishing reliable data as to the age and many other facts in the life history of that fish. It has been shown to be applicable also to various species of trout, and its value has been demonstrated in fishes as widely divergent as the carp, the eel, the bass, the flounder, and the cod. Descriptions of this scale structure and its significance have appeared in a large number of papers, both scientific and popular. It will suffice here to repeat that the scale in general persists throughout life, and grows in proportion with the rest of the fish, principally by addi-

o The life cycles of Pacific coast salmon belonging to the genus Oncorhynchus, as revealed by their otolith and scale markings. Transactions Royal Society of Canada, 1912.

AGE AT maturity OF THE PACIFIC COAST SALMON

5

tions around its border. At intervals there is produced at the growing edge a delicate ridge upon the surface of the scale, the successive ridges thus formed being concentric and subcircular in contour, each representing the outline of the scale at a certain period in its development. Many of these ridges are formed in the course of a year’s growth, the number varying so widely in different individuals and during successive years in the history of the same individual that number alone can not be depended on to determine age. For this purpose we rely upon the fact that the fish grows at widely different rates during different seasons of the year, spring-summer being a period of rapid growth and fall-winter a season when growth is greatly retarded or almost wholly arrested. During the period of rapid growth the ridges are widely separated, while during the slow growth of fall and winter the ridges are crowded closely together, forming a dense band. Thus it comes that the surface of the scale is mapped out in a definite succession of areas, a band of widely spaced rings always followed by a band of closely crowded rings, the two together constituting a single year’s growth. That irregularities occur will not be denied, and this is natural, inasmuch as growth may be checked by other causes than the purely seasonal one. Also a considerable experience is requisite for the correct interpretation in many cases, and a small residue of scales of doubtful significance has always remained. This element is too small to affect the general results, and further investigation will almost certainly eliminate the doubtful cases altogether.

SOCKEYE OR RED SALMON (Oncorhynchus nerka).

[PI. i-m; fig. 14, pi. viii; fig. 22, pi. xn.]

The sockeye, red salmon, or blueback, as it is variously known, has been assumed to mature principally if not wholly in four years, on the basis of the constant four-year periodicity in the magnitude of its run in the Fraser River. Richardson is reported as having marked young sockeye at Karluk, Alaska, and observed their return at four years on the average. Chamberlain marked young hatched from 1902 eggs and planted the marked fry in the Naha River, Alaska. The return of adults bearing the appro- priate mark was reported in 1906 and 1907, and has continued to be reported for each year since that date, including the summer of 1911. If these results are accepted, individuals may even reach the age of 9 years before maturing; but this is a conclusion so at variance with other evidence and with general probability that complete corroboration is required.

Finally, on the basis of scale and otolith structure, McMurrich announces 4 years as the age of adult Fraser River sockeye and 2 years for the undersized or grilse form. But as he frankly bases his investigation upon the a priori assumption that Fraser River sockeye must be 4 years old, and interprets scale and otolith structure in accordance with that assumption, it can hardly be claimed that his conclusion adds anything to our pre- vious knowledge. This becomes the more apparent upon an examination of his figures, which indicate that he unfortunately chose for examination certain large specimens which were in reality 5 years old and not 4. Forced by this method to make an erro- neous interpretation of the central area of the scale, all his conclusions are vitiated which

6

BULIvETIN of the bureau of fisheries

deal with the sockeye, the king salmon, and the coho. Thus the sockeye grilse is a 3-year and not a 2 -year fish, a fact which we learn with some relief, in view of the very large numbers of grilse which regularly make their appearance three years after each quad- rennial big run in the Fraser and the small numbers which can be observed in the inter- vening years. For reasons which will appear later, McMurrich’s initial error did not to the same extent affect his interpretation of the scales of the humpback and dog salmon.

Our knowledge of the life history of the young sockeye is founded first on the important researches of Commissioner J. P. Babcock for the Fraser River (Report Fish- eries Commissioner for British Columbia, 1903), and later of Chamberlain, for certain streams in Alaska. These investigators agree that in the early spring two sizes of young sockeye may be found together in the streams on their downward migration to the sea. The smaller of these comprise fry of the year hatched from eggs of the preceding fall, and are then about inches long. Those of larger size are yearlings, which instead of running out to sea the previous spring as fry, have remained in the lake during their first year and have grown during that time to a length of 3 or 4 inches.

The relative proportions of fry and yearling migrants in such streams as the Fraser and the Karluk are unknown. More fry have usually been captured in the fine-meshed nets employed for the purpose, but the superior strength and wariness of the larger yearlings have doubtless affected the result. It is also unknown whether in the same stream these two groups remain relatively constant in size or fluctuate widely from year to year in accordance with changing conditions of unknown nature. Chamberlain has pointed out that in the Naha River, Alaska, all the young remain in the lake during their first year, and migrate in their second spring as yearlings. It seems very probable that the chances of survival of yearling migrants is better than that of the fry, in which case a larger proportion of adults might be derived from that group, even in streams where considerable numbers migrate as fry. As will be shown below, the scales develop differently in the- two groups. The history of each group can therefore be traced and the proportions which attain maturity can be determined.

In his interpretation of the sockeye scale McMurrich assumes that 13 of the 14 specimens which he examined belonged to the group which migrated as fry, the central close-ringed area of the scale having been formed during their brief life in fresh water. But, as a matter of fact, at the time when young fry migrate no scales whatever have made their appearance. In this group, therefore, the adult scales contain no record of life in lake and stream, even the innermost rings having been formed under marine or estuarial conditions. The close-ringed area which forms the center of the scale in McMurrich’s specimens, as well as in the great majority of Fraser River sockeyes (pi. ii, fig. 4) , designates members of the second group, which have remained in their native stream or lake until their second spring. It presents, therefore, such record as we have of approximately the first year and a half of their life cycle. Plate iii, figure 5, represents a scale, enlarged 40 diameters, of a 5^earling 3^ inches long, taken in the Fraser River in April, 1903, by J. S. Burcham, working under instructions of Commissioner J. P. Babcock.

age at maturity of the pacific coast salmon

7

The scales in different yearlings captured at the same time and place differ considerably in size of scale and in the number of rings which they contain. Precisely similar varia- tions are found in the nuclear areas of the adult scales, these being larger in some indi- viduals than in others, and containing more numerous rings. Every variety of scale from migrating yearlings can be matched in the close-ringed centers of adult scales, both as regards actual size and in the number and arrangement of the rings. We can entertain no doubt, therefore, that the two have had an identical history. (See pi. ii, fig. 4, with center of adult scale enlarged 40 diameters.)

The peripheral part of the scale seldom offers any difficulty. After life has begun in the sea, a regular alternation occurs of bands of widely spaced and of narrowly spaced rings, as shown in plate iii, figure 6, the widely spaced rings representing the vigorous growth of spring and summer, the narrow rings the retarded growth of fall and winter. Finally, at the margin of the scale of the mature sockeye about to enter the Fraser are found a few widely spaced rings, indicating that the rapid growth of the summer in which maturity is attained early comes to an abrupt conclusion. This is true in the sockeye in greater degree than in other species of the genus, and may have its explana- tion in the earlier date at which mature sockeye discontinue feeding. All species of Pacific salmon (the steelhead is not here considered) cease to feed on entering fresh water at maturity, but the sockeye is extreme in this regard. Those bound for the Fraser are already fasting when first encountered along the Vancouver Island shore at the entrance to the Straits of Fuca, where other species are feeding greedily. The sockeyes are then over 100 miles from the mouth of the Fraser and are assuredly at that time not under the influence of fresh water. Fishermen are well acquainted with the fact that the sockeye, unlike the king salmon and the coho, are not to be taken by trolling, even when first they strike the coast.

With these facts in mind, we turn to plate ii, figures 3 and 4, which represent a scale typical of a majority of the sockeyes of the Fraser River run. The nuclear area of finely crowded rings contains no record of the first winter after the eggs are laid, but represents the first summer and second winter which were spent in the lake. The broad band of widely spaced rings surrounding the nuclear area was formed in the sea and represents the second summer, thus completing the second year of the cycle. Then follow a narrow winter band of closely spaced rings and a second distinct summer band, constituting the third year of the cycle, and another winter band and the short marginal summer band of the fourth year.

While the majority of the Fraser River run are in their fourth year as shown by the evidence here adduced, a considerable number of them, including all the larger individuals, are just as evidently in their fifth year. The 4-year fish show, outside the nuclear area, three summers’ growth and two winters’ growth in the sea. The 5-year fish, as shown in plate viii, figure 14, have just as distinctly, outside the nuclear area, four summers’ growth and three winters’ growth in the sea. As is given in a later table, there is a wide overlapping in size of the 4 and the 5 year fish, but all the smaller specimens are 4 years and all the larger are 5 years old. The specimens examined and

8

BUIvIvETIN of the bureau of fisheries

figured by Prof. McMurricli unfortunately were chosen from among the larger of those to be found on the cannery floor.

The great majority of the Fraser River sockeyes have scales of the type above described, the nuclear area being small, of crowded rings, and sharply set off from the widely spaced summer rings which surround them. As we have shown, these have all developed from fish which migrated seaward as yearlings. A different type is repre- sented by plate xii, figure 22, in which it is seen that the nuclear area is much larger, the rings less closely crowded and widening gradually outward, until in extreme cases they merge almost imperceptibly with the succeeding summer rings. A scale of this type is figured by Prof. McMurrich (pi. iii) and is interpreted by him as indicating a fish “which went to the sea as a yearling in the second spring after hatching.” But the very reverse would seem to be the case. The large size and more widely spaced rings of the nuclear area indicate that growth in those individuals which spend their first year at sea is much more rapid than in those which remain in fresh water, and this is in accord with the few experiments which have been made with king salmon to determine that point. But more conclusive evidence of the history of these fish is found on com- paring their type of scale with the scales of the humpback and dog salmon, which always migrate seaward shortly after hatching and while still in the fingerling stage. The nuclear area of humpback and dog salmon is exactly similar to the sockeye type last described, being comparatively large in size and of widely spaced rings. Further- more, the king salmon, which migrates seaward partly as fingerlings and partly as year- lings, exhibits the same two types of scale shown by the sockeye, one with a small nuclear area of crowded rings, formed as can be demonstrated during the first year in fresh water, the other similar to the “sea t3'pe” of the humpback and dog salmon. While therefore we lack such direct demonstration as could be obtained by marking sockeye fingerlings on their seaward migration and observing on their return that the scales exhibit the “sea type” of nuclear area, there is j^’et sufficient evidence for the correctness of the theory to warrant us in accepting it.

The possibility of distinguishing throughout their lives those individuals which passed to sea immediately after hatching from those which migrated as yearlings has opened up a wide field of investigation, upon which we have thus far barely entered. A certain practical difficulty is encountered at the outset. In the majority of cases there is no question to which type a given scale belongs. But among those of undoubted “sea type,” including humpbacks and dog salmon, as well as certain sockeyes and king salmon, there is found a tendency to the narrowing of a few of the rings immediately surrounding the nucleus, forming a sort of core to the nuclear area. (See pi. x, fig. 17.) The significance of this is in question, but we may perhaps hazard the conjecture that in such a case the individual tarried in fresh water or played back and forth on the tides for an appreciable time, during which growth was less rapid than in the majority which passed directly out to sea. Whatever the cause, this tendency to a slight central narrowing of rings of the nuclear area is of not infrequent occurrence in scales of the sea type, and is occasionally so pronounced as to simulate the smaller and least typical of what we may call for purposes of distinction the “stream nucleus.” In a small proportion of

AGE AT MATURITY OF THE PACIFIC COAST SALMON

9

cases this may be the source of genuine doubt as to the early history of the individual and incidentally as to its age. For, if the nuclear narrowing be interpreted as indicat- ing the first year spent in fresh water, the age will be greater by one year than if it be interpreted as the central narrowing of a large nuclear area of the “sea type.” It is believed, however, that all doubt of this character can be removed by further investi- gation. For purposes of a preliminary discussion, like the present, as the doubtful cases are few in number, they can be omitted from consideration.

There is a fair indication not to be taken as conclusive that those individuals among the sockeyes which proceed to sea shortly after attaining the free swimming stage experience a mortality far in excess of those which pass to sea as yearlings, but that those of the first class which survive grow much more rapidly the first year than they would have done had they remained in fresh water. To a certain extent they seem to maintain this preponderance in size during succeeding years. In the Fraser River, as already indicated, only a very small proportion of adult fish have developed from young which sought the sea during their first year. Thus out of 625 individuals taken at random, without selection, from the cannery floor, only 35 belong to this group, while 590 had spent their first year in their native waters. So limited a number as 35 forms a wholly inadequate basis for generalizations, but it is worthy of note that 3 of these were grilse, maturing in their third year, 30 were in their fourth year, and only 2 in their fifth. If a larger series shall verify essentially the proportions here indicated it will demonstrate that early migration of the young accelerates growth and also the early coming to maturity, producing a larger proportion of grilse which mature in their third year, and a much greater preponderance of fourth to fifth year fish than occurs among those which migrate as yearlings. The increased size at the same age becomes evident on comparing the three “sea type” grilse above mentioned with “stream” grilse taken on the same date. The former were respectively 23, 23^, 23^4 inches long and weighed 4^ to 5^ pounds. Sixteen grilse of the stream type taken on the same date (all that were secured) show the following lengths in inches: 17^, 18, 18, 18^, 18%, i8)4, 18^, 19, 19, 19^, 19^4, 19/4, 19/4, 20, 20^. Three of these, i8}4, i8}4, and 19^^ inches long weighed each 2^ pounds. It is further noticeable that 2 of the 3 grilse of the sea type were females, while among Fraser River grilse of the stream type females are so rare that we have thus far discovered but one among over a hundred grilse examined. This is a matter, however, which may be subject to wide variation in dif- ferent streams. In the Columbia River, for instance, in 1910, grilse were numerous, and males and females in approximately equal numbers.

In discussing the commercial value of a run the grilse may be omitted from con- sideration, as ordinarily they are few in number and so small as to have practically no value. The valuable elements of the run are fish which are maturing in their fourth or in their fifth years. It was attempted during the season of 1911 to determine the rela- tive proportions of these two generations, their range in size, and their average weight. In order to simplify the problem as much as possible we have omitted from our list the comparatively few individuals of sea type. In 500 fish of stream type, taken at ran- dom from the cannery floor, 271 were 4-year and 229 5-year fish. Of the 271 4-year

lO

BULLETIN OF THE BUREAU OF FISHERIES

fish, 126 were males, 145 females; of the 229 5-year fish, 129 were males, 100 females. The total number of males (255) was but slightly in excess of the total number of females (245). Among 4-year fish the females were decidedly in excess, among 5-year fish the males.

As regards weight, 113 4- year fish taken without selection aggregated 709.25 pounds with an average of 6.27; 104 5-year fish weighed 776.25 pounds, with an average of 7.46. Extremes of weight among 4-year fish were 4.75 and 8, among 5-year fish 5.75 and 8.75.

The following table gives distribution in accordance with their length of the 500 individuals of the stream type in which age was determined, the 4-year males being compared with the 5-year males, and the 4-year females with the 5-year females.® At the time the measurements were taken the jaws in the males were seldom strongly hooked, but a preorbital elongation was apparent which accounts in part but not wholly for the greater length of males than of females. The fish were measured over the curve of the body from the end of the rostrum to the tips of the middle rays of the caudal fin. It will be noted that the range in size of the 4 and 5 year fish overlap widely, but the curves are characteristically distinct, the 5-year fish averaging at least 2 inches longer, out of a total range of 5 inches in length for each group. Examination of a larger series of individuals would unquestionably extend the limits of each group by the addition of the exceptionally large and the exceptionally small, but the table as pre- sented doubtless gives correctly the range in size of the vast majority at the time the examination was made. It will be of great interest to compare similar curves prepared for the same river basin in successive years; also curves of the sockeye run in different basins.

Five Hundred Fraser River Sockeyes which had Migrated to Sea as Yearlings, Grouped

BY Sex, Age, and Size.

Length in inches.

Males.

Females.

Length in inches.

Males.

Females.

4 years.

5 years.

4 years.

5 years.

4 years.

5 years.

4 years.

S years.

21

21^

22

22M

23

23H

24

24^

25

25H

26

2654

27

21

II

4

25

22

15

IS

12

2

2

27

IS

12

4

2

I

1

2

II

19

49

27

22

II

2

4

5

13

20

23

23

28 28x4

29

29J4

Total. .

1

2

3

13

21

4

13

20

I

126

129

14s

100

Three-year fish, or grilse, were not included in the above table. During the summer of 1911 they were not numerous. Only a few (perhaps 3 or 4 on the average) could usually be found in each i ,000 fish brought in. It is the current belief among those who handle Fraser River sockeyes that all the grilse are male fish. As previously noted,

a Measurements were made August 2-4, 1911, in the cannery of the Pacific American Fisheries at Bellingham, Wash. To the management of this concern we desire to express our grateful appreciation of their uniform courtesy.

AGie AT MATURITY OF THE PACIFIC COAST SALMON II

all those examined by us were males, with one exception, a female with well-developed ova. Examination of the scales (see pi. i, fig. i) shows that the final summer’s growth is more extensive in grilse than in those which mature at 4 and 5 years, and the narrowing of the marginal rings to form the third winter band has often begun.

Much larger specimens than those included in the table were selected for examina- tion in an attempt to find individuals older than 5 years. Several were found over 30 inches long, weighing 12 pounds, but all were 5-year fish. Examination of a larger number may possibly bring to light a rare individual which has not matured until its sixth year. Even this is doubtful and we assuredly do not anticipate finding older than 6-year individuals among the Eraser River sockeyes. The significance of the mark- ing experiments of the Bureau of Fisheries in the Naha River, with the reported recovery of marked specimens up to 9 years old, remains to be determined.

If it be true, as indicated in our table, that relatively so large a proportion of Eraser River sockeyes mature in their fifth year, it may appropriately be asked how the enor- mously increased run every fourth year can be maintained in that river without its benefits becoming gradually distributed through five-year individuals over the inter- vening years. The great run of 1909 must have developed as 4-year fish from the superabundant eggs deposited by the great run of 1905. But if an almost equal propor- tion of those eggs should fail to mature until their fifth year, as was true in the 1911 run which we investigated, we should have expected a second great run, characterized by their relatively large size, in 1910. Such increased runs in the fifth years do not occur, and of this we have as yet no explanation to offer. There is some evidence that the fifth years of the cycle are characterized by fish of somewhat larger size than the other “off years,” indicating, if true, a larger proportion of 5-year fish, but the total run is not appreciably increased. Certain it is that the fish of the big runs average smaller than those of the intermediate years. This might conceivably be explained by a limited food supply and sharper competition among the enormous schools of that year, but it is more probably due to the practical elimination of 5-year fish. Those 5-year fish present would have developed in their due proportion from the few eggs of an “off year,” and would be too scattered to produce any effect among the vast hordes of 4-year olds. But these are matters for further investigation.

KING SALMON, OR CHINOOK ( Oncorhynchus tschawytscha ) .

[PI. iv-vii: fig. 15, pi. vm: pi. ix; fig. i8, pi. x; fig. 25. pi. xiv.]

Speculation concerning the age of the king salmon (also called spring salmon, tyee, Chinook salmon, Sacramento salmon) has been encouraged by the enormous range in size which is exhibited by spawning fish. Adult females have been reported as small as 5 pounds, and adult sea-run males much smaller than this, while individuals of from 80 to 100 pounds weight occasionally are seen. No answer has heretofore been given to questions concerning the total range in years represented by these various sizes, nor as to the exact relation of age and size.

A detailed experiment to determine these points was undertaken by Superintendent Hubbard of the Clackamas (Oreg.), station, in 1896. Five thousand young, hatched

12

bulletin of The bureau of fisheries

from eggs of the preceding fall (1895), were marked by removing the adipose fin, and were then liberated. In 1898“ Columbia River cannery men reported the capture of some 375 of these marked fish, indicating their return in their third year. Thirty-two are reported, with details of sex and weight. Nineteen of these were females, ranging from 10 to 35 pounds; 13 were males, from 19 to 57 pounds in weight. The average for the 32 is 27.69 pounds, about 5 pounds above the average of all Columbia River salmon of this species.

The following year ^ between 40 and 50 were reported, the average weight said to be nearly 10 pounds greater than of those taken in 1898. And in 1900, by offering a small reward for marks saved and sent in, the Oregon commissioner received 72, with no indication of weight. No further record of the capture of marked fish is contained in the Oregon reports. According to this experiment, it would appear that both males and females may mature in their third year, and that very large fish (57 pounds) may appear among those returning thus early. These results have been accepted by Rutter and others, the evidence appearing on its face unimpeachable, but they are so wholly at variance with results obtained through a study of the scales that we have scrutinized the records with some care. Certain minor inaccuracies are obvious, but the dates are well attested, and the only criticism which seems pertinent is that the commissioner apparently relied wholly on reports from the canneries, and did not personally inspect any of those specimens reported in 1898 and 1899.

Other marking experiments on the Columbia River, made by the Washington Commission at Kalama and Chinook, indicate 4 and 5 years as the age of returning fish, none being recorded in the third year. In the well-known planting experiments in Tomales Bay, Cal., and in New Zealand, both of which resulted in establishing spawn- ing runs in streams which hitherto had not possessed them, no run was reported until the fourth year. In view of these facts and others, we must hold in question the report of the Oregon experiment of a considerable return in the third year of both males and females of large size, in spite of a certain appearance of conclusiveness in the report itself.

The history of the king salmon is known principally from the work of Rutter and Scofield on the Sacramento River, and of Chamberlain in Alaska. The fact is well attested that large numbers of fry pass into salt water in spring or early summer as soon as possible after the absorption of the yolk. Many others, but in unknown proportion, remain behind in the streams and migrate the following spring as yearlings. Whether any of these, which have been somewhat unfortunately termed “summer residents,” pass out during the late summer or on the high water of the fall remains unknown, but many if not all of them remain during the following winter. The center of the king salmon scale, which records their early history, is subject to more variation than in the sockeye, and will require longer investigation in connection with young salmon of known history.

o Sixth Annual Report Fish and Game Protector of State of Oregon, 1898, p. 48. b Annual Report Department of Fisheries, State of Oregon, for 1899 (1900), p. 15.

AGE AT MATURITY OE THE PACIFIC COAST SALMON

13

In general, king salmon scales exhibit the same two types characteristic of the sockeye a stream type, with close-ringed nuclear area, sharply set off from the envelop- ing wide-ringed summer band (pi. iv, fig. 7), and a sea type, of large nuclear area, with less crowded rings, which widen outwardly and usually pass gradually into the rings of the second summer (pi. vii, fig. 13). The nuclear area of the close-ringed stream type agrees with the entire scale' of a yearling on its seaward migration in its second spring, and can be safely so interpreted. The nuclear area of the sea type is here interpreted as in the sockeye, as indicative of an early passing to salt water on the part of those individuals which migrate as young fry. A more extended inquiry into this matter must be made before offering data with full conclusiveness, but it may be offered in evidence that young king salmon about 4 inches long taken in Puget Sound in mid- summer exhibited the wide-ringed sea type and were in their first year (pi. vii, fig. 12). It may also be noted that on examining a series of larger king salmon (9^ to 17^ inches long) taken in late summer in Puget Sound, all were found to be in their second year, the smaller individuals (9>2 to ii>2 inches) exhibiting the stream type of nucleus (pi. iv, fig. 7), and the larger members of the series (13K to i7>^ inches) the sea type (pi. vii,

fig- 13)-

Conspicuous in every spawning run of king salmon are the numerous undersized males, known locally as grilse, jack salmon, or sachems. Two theories have been held regarding these, according to which they have been considered either stunted individuals of equal age with the larger salmon, or younger fish which have matured precociously. The theory of precocious development has had wider currency of late, and is in entire agreement with the evidence from the scales, according to which the individuals are always in their second or third year. The mature second-year fish are smaller than those in their third year, and are usually little in evidence, as they escape readily through the meshes of the nets. None which were mature at this age have been taken by us in Puget Sound. But on the Columbia River, a considerable series of mature males in their second year, 9 to i8}4 inches long, were secured from the seines and fish wheels. It will be noted that this range in size agrees with that already given for immature second-year fish from Puget Sound. It is further significant that in the Columbia River series, also, the smaller individuals, 9 and inches long, are of the stream type, and the larger, 13 to i8>^ inches, are of the sea type.

The larger grilse (19 to 26 inches) are in their third year (see pi. iv, fig. 8; pi. vii, fig. ii), and among them, again, the smaller individuals of the series are preponderat- ingly fish of stream type and the larger of sea type. But in the third-year fish there is more overlapping of the two types, as though subsequent inequalities in growth had partially concealed the initial advantage secured by those which had early sought the sea. This is a question which merits further investigation on a much larger series than has thus far been examined.

From salt water in Puget Sound we have secured immature third-year fish, both males and females, and also matured third-year males, taken by purse seines from the same school, and both feeding voraciously and equally on small sand lance and young 60289°- Bull. 32 14 2

14

BUI^LETIN OF THE BUREAU OF FISHERIES

herring. There was no difference in size between the mature and the immature indi- viduals, nor could they be externally distinguished, unless by a certain distention of the abdomen in mature specimens, due to the developed testes. It became evident from our observations: (i) That a very small proportion of the males of a given year develop precociously; (2) that precocity is apparently not caused by the influence of peculiar external conditions operating upon the individuals thus affected, but by some unknown factor; (3) that precocious development does not stunt the growth. No mature female king salmon less than 4 years old have thus far been encountered.

The commercially valuable portion of the king salmon run consists mainly of 4 and 5 year fish, with less frequent 6-year individuals. Plate v, figure 9; plate x, figure 18; and plate vi, figure 10, represent these three ages. On the Columbia River they are roughly grouped at the canneries as “half salmon” and “full salmon.” The half salmon consist very largely of 4-year individuals and the full salmon those 5 and 6 years old, although there is a certain amount of overlapping, as in the sockeye. The 4-year fish include more females than males, and the