CONFrDSKTIAL v

FISHERIES ("ADA

STMiaRY K£P0RT3

of the

PACIFIC BIOLOGICAL STATION

for

1940

.0, B.C DUCiZdMt, l°4u \ '<:••/-'

FISHERIES RESEARCH BOARD OF CANADA

INDEX TO SUMM&RY REPORTS FOR 1940

5s-

Salmon Investigations: Pink salmon Nos. 1 - 11 • Page 1-15 Chum salmon 12 - 13 16-18 Sockeye salmon • 14 19 Spring salmon 15 - 17 20 - 23 Coho salmon 18 - 19 24 - 26

Cowichan River Investigations: Salmon 20 27-28 Trout 21 - 22 28-30 Hatchery and limnology 23 - 25 31 - 34 iV^ Predators and competitors ^x^ 26 34 - 35 h

Ocean Fisheries Investigations: Herring 27 - 44 36 - 61 Pilchard 45-48 62-66 Ling cod 49 - 50 67 - 69 Smelt 51 - 53 70 :r- 73

Shellfish Investigations: Oyster 54 - 56 74 - 76 Clam 57 - 62 77-80 Miscellaneous 63 80

•Hydrographical Investigations: General 64 - 65 81 - 82 Alberni inlet 66 82 - 83 Physical oceanography and salmon 67 - 68 83 - 85

Fish Diseases Investigation: Furunculosi s 69 86 - 92

Other Investigations: Identification of Polychaete worias 70 93 Economic relation of water fowl 71 94 - 95

\rM.

>v/> -1 -

A*L. Pritchard No# ^

Physical Characteristics of the Adult Pink Salmon at McClinton Creek, , in 194-0.

As one phase in the investigation of the natural propagation of pink salmon at McClinton creek, Masset inlet, close examination of such characteristics as size of the individuals and the egg content of the females, has always been carried out. During the run of 1940, co incidental with the tagging programmes described in another report, 983 males and 840 females were taken at random, measured to the nearest tenth-inch, and weighed to the nearest tenth-pound. Seventy females were killed and the eggs in each accurately counted. The male pink salmon ranged in length from 18.3 to 26.0 inches, averaging 21.9 +, #025. The females, as is apparently usual, were dis tinctly smaller7 varying from 17.6 to 22.8 inches, with an average of 20,2 .+.015. These are the smallest mean lengths recorded for the last six cycles with the exception of 1930 when they were: males, 21.125 + .029, and females, 20.125 _+ ,023. "" A similar difference existed in the weight of the two sexes. The males varied from 2.5 to 7.6 pounds, averaging 4.2 + .016. The fe males ranged from 1.9 to 5.0 pounds, with an average of~3.6 + .009. In 1930 only were lower averages obtained. ~" In order to ascertain whether there was a change in the size of the individuals with the progression of the run, the sample was averaged in quarters, the first quarter being first in point of time of migration. For both sexes there was a significant change in length, the fish at the end of the run being larger. There is no doubt that some of this increase in the case of the males was due to the development of the long snout and head as secondary sexual characters. This ex planation will not suffice to explain the increase in the case of the females which are known to change little as the result of sexual matur ity. Peculiarly enough, the variation in weight was in the opposite direction. The females taken at the last of the run were significantly lighter than those taken at the beginning. The difference between the averages for the first and last quarter in the males was just on the border-line of significance. This condition may be attributable to a peculiar environmental circumstance which occurred during the season. In the first part of August heavy freshets took place. The result was that the pink salmon immediately migrated upstream. These fish had ap parently come in almost directly from the open sea and had not remained long off the river mouth since none were observed schooling in the bay, and it was extremely difficult to differentiate between the sexes by ex ternal examination. It can be assumed, therefore, that these migrants had not long ceased feeding. They would thus not have lost so much weight as those which migrated at the end of the run in a well developed sexual condition. That the 70 females used for egg counts might be considered typical was shown by the fact that their mean length and mean weight, 20.3 _+ f056 inches, and 3.6 £ .035 pounds respectively, did not differ significantly from the mean of all the females recorded above. The aver- - 2 -

A,L. Pritchard No. 1 age number of eggs per female was 1,619 £ 17. "This is the smallest figure obtained with the exception of 1930. A direct significant cor-' relation existed between the number of eggs and each of the length and weight. The calculated coefficients were: #45 £.048 and .77 £ .032 respectively.

A.L. Pritchard No. 2

Natural Propagation of Pink Salmon in McClinton Creek, Masset Inlet, B.C. - The Mature Salmon Run in tho Autumn of 1940. . j j_ . i ti 1_ I jI in [ I -- 11-1 — T ^^—. The following report is a summary of the general observations and counts on the natural run of adult pink salmon to McClinton creek, Masset inlet, in the autumn of 1940, For purposes of comparison a table embodying the numerical data for this and previous migrations in 1930, 1932, 1934-, 1936, and 1938 has been set up. The period of upstream migration of pink salmon into McClinton creek in 1940 was more extended than usual, beginning considerably earlier and continuing somewhat later. Tho first fish appeared at the fence on July 28. Thereafter until August 3 only scatterod individuals arrived daily. From August 3 to 14 the numbers amounted to over fifty per day to produce a total of approximately one thousand. The count then decreased and remained small until August 26. The main run occurred be tween August 26 and October 7, After the latter date there was an abrupt decline. The largest daily migration was on September 19, when slightly over 5,900 individuals were handled. The numbers of pink salmon reaching the weir were again defin itely correlated with rainfall and the consequent changes in the height of the water in the creek. The migration of over one thousand fish during the heavy rains from August 3 to 14 was especially striking since these salmon were so fresh that it was difficult to differentiate males from females by external examination. Such a condition is unusual. Individuals of both sexes made their appearance at almost the same time. The males, however, increased relatively until September 14. For the next week proportionately more females arrived. The final dispo sition of the sexes on the spawning grounds was in the end significantly unequal - 53.7 per cent males, and 46.3 per cent females. No evident explanation for this difference is available at present. Using the average number of eggs per female reported elsewhere (1,619 £17), the potential egg deposition was calculated as 26,500,000. The following table presents a summary of the counts obtained in the natural propagation experiment to date: - 3 -

A.L, Pritchard No. 2

.<6". Potential Egg Fry i Year Males Females Total No. of Eggs Deposition Migrants Hatch

1930 32,955 33,198 66,153 1,535 £ 12 50,950,000 5,384,000 10,6 1932 8,003 7,597 15,600 1,758 £ 15 13,360,000 2,230,000 16.7 1934 77,477 77,719 155,196 1,799 £ U 139,000,000 12,600,000 9 a 1936 24,221 28,091 52,312 1,899 £ 12 53,345,000 3,675,000 6.9 1938 5,549 5,028 10,577 1,698 £ 19 8,50(5,000 2,020,000 23.8

1940 19,071 16,454 35,525 1,619£ 17 26,500,000 - -

In summary reports of previous years it has been suggestod that a low or high efficiency of hatch will be carried through the life cycle to a greater or less extent and give rise to a small or largo adult run respectively. If this suggestion is sound, a relatively large migration .2>> should have occurred at McClinton creek in the autumn of 1940 following the top efficiency of 23.8 per cent. The actual returns were, however, of only medium size. In spite of this apparent refutation of the assum ption, it should not be discarded entirely until the complete detail of the recovery of marked McClinton pinks is worked over. This may indicate that there was abnormal exploitation of tho run before it reached the creek* The return in fry migrants from the seeding of 26,500,000 eggs will be determined in the spring of 1941. Since this number of eggs is quite different from the other seedings, opportunity is thus afforded to collect information for another set of conditions. Valuable data should result.

A.L. Pritchard No. 3

Recovery of Marked Pink Salmon in 1940.

In connection with the investigation of the natural propagation of pink salmon being carried out at McClinton creek, Masset inlet, it was desirous to have more factual data regarding the behaviour and migrations of the species during the ocean phase of its life history. Accordingly, in the spring of 1939, 178,629 fry migrants out of the total of 2,020,000 leav.ing the stream (8.8 jo) were marked by the removal of both ventral fins. These, according to indications from previous experiments, were expected to appear in the fishery during the autumn of 1940. Every possible offort was made to locate the scarred individuals< An observer was placed in each of certain representative canneries whose sole duty it was to examine all the catches coming into the plant. The following areas were chosen: Arrandale cannery on the Nass river, Prince Rupert cannery in Prince Rupert harbour, Sunnyside and Claxton (part time) canneries on the , Butedale cannery in the central northern, district, Goose Bay cannery in , Bones Bay cannery (part time), Pacofi cannery on the southern end of the Queen Charlotte islands, - 4 -

A.L, Pritchard No. 3 and Shannon Bay cannery in Masset inlet. In addition, the United States Fish and Wildlife Service were kind enough to assume full res ponsibility, financial and otherwise, for a collector at Waterfall on Prince of Wales island in south-eastern Alaska. In those areas not covered by examiners, wide advertisement was given to the fact that the marked pink salmon were expected. Many of the rivers tributary to Mas- set inlet were inspected, and the whole migration to McClinton creek, where the fish had originated, was counted and examined. Complete data for all the collections are not yet available, but those already received are presented in the table below. There un doubtedly will be slight changes in some of tho figures, but these should not affect any of the general statements made herein.

No. Fish No. Area of Examination Examined Marked Location of Capture

COMMERCIAL CATCHES:

Waterfall, south- eastern Alaska 2,000,000 30 Off Noyes, Baker, and Dall is., south eastern Alaska Nass river 50,000 Prince Rupert Not reported Skoena river 200,000 1 Off mouth of Skoena 1 Skeena river Butedale 175,000 0 17,000 8 Masset inlet Goose bay Not listed 0 Bones bay Not listed 0 Masset. inlet 175,000 142 Masset inlet (exclu sive of McClinton bay). 6,000 n McClinton bay Pacafi 250,000 0

SPAJWNING STREAMS:

Masset inlet Yakoun river 1,647 0 Mammon river 3,400 1 Detlamen river 400 1 Deenan river 275 0 McClinton creek 35,525 791

One striking feature recorded by the observer at Butedale was the marked contrast in the proportion of scars from fish caught in Masset inlet and those taken in other areas. In 175,000 fish from localities other than

y -5 ~

A.L. Pritchard No. 3

Masset not a single marked individual was found, yet in only 17,000 Masset fish, eight were easily located. In Masset inlet itself a simi lar contrast was demonstrable between McClinton bay and the rest of the fishing areas. It should not be too much to conclude, therefore, that a greater proportion of marked pink salmon found their way into Masset inlet and thence to McClinton bay than to any other district. The consideration of the recoveries in McClinton creek itself shows a return of only 2.2 per eent. At the time of fry migration tho ratio of marked to unmarked individuals was 8,8 per cent. This discrep ancy still has to be explained, although several suggestions as to the cause of the difference are submitted below. The examination of certain other rivers in Masset led to the discovery of McClinton marked fish in at least two. It must therefore bo admitted that some wandering did occur. If this is the case, it is not illogical to assume that some unmarked pinks found their way to Mc Clinton. This exchango would have the effect of decreasing the percent age ratio of marked to unmarked amongst the adults. There is also little doubt that some differential mortality exists, since it is difficult to believe that the removal of fins from small fry will not harm any of them. To date no effort has been made to determine the absolute amount of effect of each of these factors. The returns from Alaska are interesting, important and grati fying. From the point of view of the fishery it demonstrates definitely that the Alaskan fishermen may exploit the Canadian runs in general and the Masset run in particular. There is no doubt that the amount of this exploitation will vary from year to year depending on the route of migra tion of the fish and the intensity of the fishery. Thus the runs to the inlet may fluctuate in size because of factors beyond the control of the Canadians to regulate. There seems to be little doubt but that work of this type should be carried out again during the coming spring. When the returns are ex pected every effort should be made to increase the search for returns in south-eastern Alaska so that full data may be available on this important phase in the migration. The arrangements should be easy in view of the excellent and unstinting cooperation which was given this year.

A.L. Pritchard and W.M. Cameron N0# 4.

Variations in the Physical and Chemical Conditions in McClinton Creek, Masset Inlet, during the Spawning Runs of Pink Salmon.

The very definite and significant correlation of the numbers of spawning pink salmon migrating upstream in McClinton creek with the rainfall in the area and the height of water in the creek has suggested that a close study of the variation in other physical and chemical condi tions might illuminate some of the basic factors which determine the mi gration of these fish. During the season just passed Mr. H.J. Hollister carried out a large number of observations, the results of which are here briefly summarized. - 6 -

A.L, Pritchard and VJ.M. Cameron No. 4

As a result of data derived from a large number of determina tions of dissolved oxygen by tho Winkler method during tho spawning mi gration of 1938, the writers came to tho conclusion*that there was a daily variation under normal conditions in clear weather, the oxygen content rising to a maximum shortly after midday and falling to a minimum during the hours of darkness. In rainy weather which was accompanied by a marked rise in the creek water level, there was apparently a defin ite interference with this rhythm. Such conditions were almost immedi ately reflected in a sharp rise in the oxygen content of the water, such a rise being more apparent when tho rain occurred in the afternoon or evoning, at which time the oxygen content would normally be on tho de cline. A rain before noon tended to accentuate the rise and result in a higher maximum for the day. There was some evidence that when a quick hard rain of short duration took place, the oxygen made a quick rise, but soon fell back to the approximate value ordinarily obtaining on a cloudy day at the corresponding time. When the rain was of long duration and the creek maintained a high level, the oxygon content of the water re mained high. When rainfall occurred in a series of short hard rains, the oxygen values fluctuated in the neighbourhood of the maximum. The data derived from the much more comprehensive series of oxygen determinations made throughout the spawning run of 1940 at McClinton creek have almost wholly confirmed the findings which have boon outlined above. For the accurate determination of the changes in the hydrogen ion concentration of the water in the creek throughout the run, it was planned to use a Beckman pH meter. Due to technical difficulties this instrument could not be employed. It was thus necessary to try to get an approximate measure of the conditions through readings with La Motte in dicators and standards, Readings obtained in this manner should be re garded with caution, since the creek water apparently presented very little buffering capacity. They show, however, that in periods of low water the pH was between 6,8 and 7,0. During freshet conditions it fell to approximately 6.5 and gradually rose to its previous value as tho water regained its normal level. The determination of the carbon dioxide tension from the pH read ings alone is impossible, but it is not unreasonable to assume that there was a change in the carbon dioxide concentration just as there was in the pH, Previously the upstream migration of salmon has been shown to be influenced by changes in rainfall, water height, current velocity, and river discharge. Since there appears to be a coincident variation in oxygen and carbon dioxide concentration , there is a possibility that the changes in these factors brought about by varying water conditions may bo the primary stimuli rather than the actual water conditions themselves. It has been observed at McClinton creek that, in the absence of freshet conditions, pink salmon moving up in any twenty-four hours enter the pens largely at night. Though this may be due to the darkness, it should be noted that at this time the carbon dioxide would be at its maximum level. Similarly, at Williams creek, Lakelse lake, where the correlation of mi gration with rainfall was not so marked, upstream movement largely ceased - 7 -

A.L. Pritchard and W.M. Cameron No, 4 in the middle of the day and a surge of fish into the pens occurred in the late afternoon. It was felt that this phenomenon was not entirely due to tho diminution in tho intensity of light. Since there is some evidence that carbon dioxide tension may play an important role in controlling salmon migrations, the advisability of a thorough and accurate measurement of the change in this character istic throughout the spawning run should hot be overlooked.

A,L, Pritchard and W.M. Cameron No, 5

The Distribution of Spawning Pink Salmon in McClinton Creek. Masset Inlet. " The general principle of the type of tagging which was developed to permit identification of individual fish on the spawning grounds, was first described in detail in the summary reports for 1938,. Briefly it involved tho use of different sized and variously coloured celluloid discs which were affixed to areas in the body by means of a nickel pin. In the autumn of 1939 in work on sockeye salmon at Lakelse lake, certain modifications were introduced, Different and more easily discernible colours were used and the areas of attachment wore limited to two, viz, in front of the dorsal fin and immediately under tho dorsal fin. Fur ther beneficial changes were introduced at McClinton creek during the season just passed. By using three layers of celluloid on each side of the fish and a variety of shapes, the individual numbers could be deter mined from either side. A few additional minor modifications should re sult in a definitely satisfactory system. An effort was ma.de to tag each day five r>er cent of the adult pink salmon migrating into McClinton creek. Occasionally when a heavy run occurred in the late afternoon or evening, tagging could not be accom plished until the following day at which time enough were marked to bring the total tagged migrants to the desired percentage. Just at the end of the run slightly too many were tagged, which raised the proportions above the amount fixed.. The experiment involved in all 983 males and 840 fe males. Tho creek was surveyed as often as weather conditions would permit, an entire inspection taking from two to three days. To facil itate observation and recording, white boards, approximately eighteen inches square and bearing consecutive numerals, were set up every one hundred yards along the length of the creek; The location of the fish was noted to the nearest one hundred feet. Generally one-third to one- half of the tagged fish were observed on any one trip. The remainder were apparently in deep pools, under jams, etc., where they could not bo dis covered, • The complete data obtained during the run of 1940 have not yet been treated in an exhaustive manner. Before final conclusions can be drawn, the figures must be submitted to rigorous statistical tests, some - 8 -

A#L. Pritchard and W.M. Cameron No. 5

of which have yet to be modified or developed. Certain general deduc tions may be drawn from a brief survey of the data. Migration of the adult, pink salmon in tho creek is intimately bound up with freshet conditions and is apparently influenced by the relative sexual maturity of the migrants. Fish passing through the fence immediately before or during a rise in tho creek level proceed prepon- derately to the upper reaches. Those individuals tagged after the peak of the freshet congregato in the lower half of the stream. A later freshet causes another wave of migrants through the fence to a position far upstream, together with a similar surge of those which are approach ing maturity in the holes immediately above the fence. These latter are replaced by frosher fish which pass the weir during the drop in water level0 It is suggested that the intensity of the reaction to the stimulus coincident with high water bocomes greater as sexual maturity approaches, but caases once spawning has commenced. Between periods of high water, the population in any ono area remains relatively static ex cept for a small degree of individual movement. During this time the fish apparently move into adjacent shallows and spawn there. It is possible that a similar wave-like migration typified the spawning run of sockeye salmon into Williams creek, Lakelae lake, in 1939. Observations there suggested a movement somewhat similar to that described, but due to the difficulty of obtaining comprehensive data and to the comparatively short time the f^sh remained alive in the stream, tho evidence could not be considered statistically reliable. The migration of pink salmon through the fence at McClinton creek has been shown by the senior author to be directly correlated v/ith the rainfall in the district and the water height in the stream. The evidence obtained this year effectively demonstrates that not only the migration past a given point but also the extent of migration and the ultimate distribution of these fish is related to those conditions.

A.L. Pritchard and W.M. Cameron NC# £

The Estimation of a Spawning Population of Salmon by the Use of a Tagging Ratio.

The estimation of the size of an animal population by tagging and releasing a known number and considering the proportion of marked and unmarked individuals in subsequent recaptures is a familiar practice, and considerable experimentation has already been carried out to define the conditions necessary for an accurate determination. The use of such a method to determine the size of spaxvning populations of salmon is re ceiving considerable attention at the present time. There are several essential facts which must be borne in mind when the system is applied. In actual practice there can be no knowledge of the relation which the number tagged bears to the total population - 9 -

A.L. Pritchard and W.M. Cameron No, 6

being investigated, since the sizo of the latter is unknown. In tho case of salmon runs, therefore, the investigators must be sure that in • examining the spawning fish (the recaptures) for the determining ratio, they are seeing a random sample of the whole run. In some instances this is possible, since the migrations are small and the spawning areas are easy of access. Practically every fish may thus be observed. If, however, only a small proportion of tho fish may later bo inspected, the chances of error become exceedingly great, especially if the distri bution of the individuals on tho spawning beds is not random. During the past autumn the writers have been attempting to test the applicability of the system at McClinton creok where thojrjiage, tho advantage of knowing the actual size of the total run, f^erean ef fort was made to tag five per cent of the daily upstream migration, Oc* easionally when a heavy migration occurred in the late afternoon or early evening tagging was impossible until the following day when sufficient were tagged to bring the total to tho required percentage. The count of dead fish after spawning has been suggested and would certainly appear to be the most practical method of discovering the tagged proportion in the stream. Accordingly, regular and thorough inspections and counts were made of the whole length of the spawning area. Sevoral complexities have been discovered which are here outlined. In an accompanying report marked differences have been demonst* rated in the dispersal of live fish which passed the fence on successive days. Comprehensive surveys of the location of the dead fish in the stream have shown that these variations were real and remained permanent. As an examplo, the fish which were tagged on the day of the freshet of September 18 wero found dead a mean distance of 2,050 yards above the •/ fence. Those tagged on September 19 were grouped around a mean distance of 1,250 yards. The difference between these means is significant, being over four times the standard deviation. The effect of the unavoidable lag in tagging occasioned by high water was strikingly brought out by the discovery of a very low pro portion of tagged fish in the upper half of the stream, Tho upper limit of the tagged percentage in this area was 2.3, using a confidence co efficient of .99 in a Polsson frequency distribution. Its discrepancy from 5.00 per cent, the actual percentage tagged at the fence, could not be balanced by that found in the lower half of the stream, viz. - 4,83, The dead fish which washed down against tho fence presented further evidence that the distribution was far from random. Those showed not only a significantly greater proportion of pink salmon which had been tagged on days immediately following high water, but also were made up largely of individuals tagged after the middle of the run. This suggests that they had not proceeded far upstream before spawning, ^___ ——__ An effort was also made this year to duplicate the method of tagging which would be necessary in an unfenced stream where a definite proportion of the run could not be assessed. Fish were dipped out of three holes approximately one hundred, two hundred, and five hundred yards below the fence. Though the analysis of the evidence is not yet com plete, the nature of the movement indicates that the fish from the lowest - 10 -

A.L, Pritchard and W.M. Cameron No. 6

hole were tho last to proceed upstream. It was noted also that the population of this area consisted of a majority of females, c. condition which usually exists at tho end of a spawning run. Another pertinent observation was that these tagged fish which after passing the fonco remained for considerable time in the lower reaches, pushed upstream on the next freshet and like those tagged on the rise or peak of high water, tended to concentrate in the upper half of the creok. Segrega tion of this type would prevent tagging in a proportionate manner. It was shown at Williams creek, Lakelse lake, in 1939, that unless the tags were applied in a constant proportion throughout the run a reliable es timate was apparontly not possible. It should be pointed out that evidence from the recovery of marked fish in Masset inlet shows that some of the pink salmon which will eventually spawn in McClinton creek do not reach it until after a con siderable number have already migrated upstream. It would seem then that since for the application of this method it appears imperative that a constant proportion of each day's migrants must bear tags, its use in a stream similar to McClinton will be exceedingly difficult, if not de cidedly doubtful.

W.M. Cameron No, 7

Further Observations on the Spawning Habits of the Pink Salmon.

In a summary report for 1938 tho spawning habits of pink sal mon at McClinton creek were described insofar as thoy could be deter mined from inspections made that autumn. Certain additional details have boen observed this year. Theso are recorded herein. It has now beon definitely established that a given female may construct more than one nest, A tagged fish observed constructing a redd on September 3 was engaged in a similar activity on September 6, but at a position about fifty feet from the original location. Due to the density of spawning in the area it could not be accurately deter mined whether she had deposited eggs in both places, Such movement would provide a reasonable explanation for the fact that in digging for eggs it is seldom possible to obtain from one redd a number comparable to the total deposition of an average female. Very few investigators have seen the spawning act of salmon or trout and no report has yet been published which describe a female spawning with two males. For this reason the following description is reported. A female had passed tho counting fence on September 19, and, when seen ten days later, had taken up her position approximately two miles upstream near the limit of the spawning area. Here the gravel was fairly coauso with a preponderance of stones two to three inches in di ameter. The excavation which the fish had made was shallow, Tho water flowing over it was six to eight inches deep and moving briskly. The observer's first approach frightened tho female into the - 11 -

W.M, Camoron No, 7

deeper water under an eidjacont bank, but soon she swam nonchalantly over to hor rodd and took up a position in its upstream half, hor vent lying two or three inches above the deepest point. Here she rested for a minuto or so, during which time sho was approached by two males. As they assumed their positions on either side of her she began to writhe hor body in a manner which can only be described as sinuous. The males wore at a slight angle to her body; tho throe tails were pressed tightly to gether; the female's head swung in a thirty degree arc between those of the males, touching first one and thon the other. A few seconds later the three mouths opened widely; a quick spurt of milt appeared and floated downstream in a brown cloud of sand. The action was practically instantaneous; the milt disappeared in tho space of a few seconds. The female pink then moved forward a foot, and with short quick spasms began to throw gravel and sand over her deposited eggs. The males began to jostle each other and were soon approached by six others who strove to usurp thoir positions. The female paid them no at tention merely carrying on by occasionally stirring up tho gravel in that characteristic manner which seoms to differentiate this phase of her activity from that of excavation. During the next hour no further spawning occurred. The female often rushed at others of her sex who apparently were encroaching on her territory. If frightened by the ap proach of the observer she would dash away but return within a few mom ents to continue hor activity. It was ascertained that forty or fifty eggs had been expelled during the act. These wero grouped closely together just below where the vent of the female had been. They were still soft and flaccid and adhered closely to the stones in the depression.

W,M. Cameron ^0 g

Studios on the Incubation of Pink Salmon Eggs at McClinton Creek in 1940, • ——

An investigation of the losses occurring during the incubation of pink salmon eggs in McClinton creek is receiving close attention at the present time. Studies of the spawning of 1938 demonstrated the •necessity for thorough and continuous sampling of the rodds to obtain reliable information. The collection of comprehensive material is thus receiving first consideration. Since this operation is time-consuming, the majority of the samples collected thus far have only been preserved and stored until the opportunity is presented for complete analysis. For this reason the following report is of necessity more general than might be desired. Preparatory to sampling, a large number of redds located in all types of bottom and water conditions wero carefully marked. Two stakes driven firmly into an adjacent bank indicated the direction of the marked redd. The distance from one stake was measured in feet and inches. The approximate shape and aroa of the disturbed gravel was sketched in a noto book and pertinent information regarding its external condition recorded. - 12 -

W.M. Cameron No, 8

In sampling, marked redds are chosen at random before proceed ing to the spawning areas. If the redd chosen proves unworkable because of high water, it is sampled at the first opportunity after that date. In.contrast to tho procedure in the previous study, the entire volume of the redd is turned over. Digging is restricted to the top six inches until that stratum has been entirely removed. Any eggs found therein are preserved separately if they show a significantly different mortal ity from that.encountered in the main body at a greater depth. At first digging is confined to a small area. As the eggs in a particular pocket are exhausted, they are removed from the net before the operations are extended, and are sorted immediately into live and dead. In this way as many as eight lots are collected from one redd, representing the sur face sample and three other parts of the deposition. Such sampling by parts cannot bo accomplished in every nest excavated, but it is attempted whenever time and circumstances permit. The eggs are then carried to the laboratory, treated and preserved with tho reagents which have been de vised to suit their particular condition, and stored for later study. Occasionally, when a redd is opened, a condition of very high or very low mortality will be evident. In such cases, especially if the sample appears to be homogeneous throughout, two or three hundred eggs are counted, placed in cloth bags of sufficiently wide mesh to per mit ample circulation, and buried in the exact location from which they were obtained. Perhaps the new situation of such eggs may hardly be normal and perhaps the excavation may have changed the character of the gravel, but it is hoped that later study of the plantings will at least give an indication of the history of the sample. As was to be expected in the larger run this year, the redds are more concentrated than during 1938-39. To date the greater propor tion of the samples have been found to represent more than one deposi tion. This condition, though complicating the study, and making diffi cult the allocation of the dead, is already suggesting that a hetero geneous sample is associated with somewhat heavier mortality. Whether this is due to the .disturbance of developing eggs by females spawning later in the same locality, or to a shifting of the over-lying gravel during high water, is not yet apparent. Later study may show it to be only coincidental. The investigator is attempting to study the effect of the de composition of dead eggs on the accuracy of the sample. To date no dif ference has been discovered in the relative recovery of live and dead eggs. All developing eggs are now well eyed, and in very recent samples a few deaths after eyeing have been noted. In order to obtain an approximate figure to present in this report, twelve of the last samples have been counted and examined. The percentage dead before eyeing in each has been calculated. The distribution of the percentage mortality for these roughly approximates a skewed bimodal frequency curve, ranging from 2.6 to 48,8 per cent and giving a mean percentage mortality of 12,1, Using the approximation of a normal frequency formula, the probable error of this figure is 2.6 per cent. Material is being collected for embryological studies in the fu ture, procedures and apparatus are being steadily improved, and it is con fidently hoped that valuable information will be forthcoming at the com pletion of the winter's investigation. - 13 -

A.L, Pritchard and W.M. Cameron No. 9

Some Physical Characteristics of Salmon Eggs and Their Relation to Redd Sampling,

Since the inception of the salmon redd sampling programme a number of refinements havo been made in both the apparatus and the pro cedure. Various nets have been tested and more satisfactory types have been developed to meet specific conditions. In this way the efficiency of the recovery of the dislodged eggs has been increased, A certain amount of experimentation has been necessary to assist in tho correct interpretation of the material so recovered. The work is continuing at the present time, but it v/ould seom advantageous to report on several experiments briefly to illustrate the progress made to date. Tests performed at Cowichan lake on coho salmon eggs dealt with the interpretation of the shells and egg fragments which usually make up a small proportion of every sample. The findings are discussed below under two headings: 1, The source of the whole shells in the sample. The break age of live or translucent eggs results in the appearance of their shells in the collecting net. These are typically clean, translucent, and ex hibit a single split seldom embracing more than half the circumference. These shells may represent either live eggs or infertile eggs which have not become opaque. It has thus been necessary to determine the relative breakage pressure of both types. As a result of crushing samples of each on a spting scale it has been found that there is no significant differ ence in the resistance of the "chorion" to rupture. Individual variation is great; some eggs of both types will successfully resist a pressure of fifty to sixty pounds per square inch, while others fracture at much lower pressures. The proportionate recovery of the two types of eggs under nat ural conditions has been investigated by burying them together and dig ging them up in the normal way. In four tests the discrepancy between the actual known ratio of fertile to infertile and that calculated from the whole eggs recovered favoured the fertile eggs very slightly. In the proportions known to exist in nature this difference would have no noticeable effect on the figure already given for the efficiency of fer tilization. 2. The recovery of disintegrating eggs. Dead eggs are known to break at much lower pressures than live ones. Due to the coagulation of the yolk and subsequent infection by fungus in many casos the fragments or shells are noticeably different from those of live eggs and can bo so allotted. The lower recovery as a result of the disintegration may, how ever, be an important source of error. Experiments similar to those out lined in the section above have shown that when approximately half of the dead eggs are broken, the recovery is about ninety per cent. When most or all are broken,the recovery drops as low as sixty per cent. This question is now receiving close attention. At the completion of the pre sent experiments a reliable estimate of the effect of disintegration should be available for each type of sample obtained. - 14 -

W,M. Cameron No. 10

Experiments on "Marking" Salmon Fry in Order to Assess the Effect of Predators.

In a summary report for 1939 it was suggested that the "marking" and release of pink salmon fry in McClinton creek at various distances above the counting weir and the determination of the propor tionate number of marked fish recovered subsequently at the structure, offered a possible method of assessing the effect of predators on the fry migration. For the practical application of such an idea a "mark" of distinction had to be developed which would* allow the easy and quick recognition of the survivors amongst the large numbers of other migrants. Experiments were performed in the spring of 1940 with chum salmon fry raised at the Cowichan lake hatchery to discover the best method of distinguishing the young fish for this purpose. The general lines of investigation are outlined briefly below: Intra-vitam staining. Bismark brown, neutral red, and nile blue sulphate in varying strengths were used to stain the fish in bulk. Staining sufficiently dense to permit identification twenty-four hours later proved fatal within two or three days. Surface staining. Most of the common stains, dyes, and inks were applied individually to various parts of the fry. The colouring washed out quite rapidly in water. The manipulation was tedious and much too slow. Branding and scarring. The application to a small area of the fish of chemicals ordinarily used as fixatives, and branding with a hot needle generally resulted in death. nThreading". The most successful method of distinction insofar as the lack of effect on the survival and activities of the fry were concerned was that in which a short length of silk was inserted through the back just below the skin. The thread is very easily and quickly seen, a large variety of colours and thus combinations are available, and it is relatively inexpensive. Dexterity in applying the thread will no doubt be quickly developed. It may be possible by this means not only to estimate the loss due to predators, but also to discover whether there is a differential mortality due to marking fry by the removal of fins. Various authorities have suggested that, though there may not be any fatal effect in removing the fins, the absence of one or more of these appendages may make the fish more vulnerable to its enemies. The experiments will be tried under natural conditions during the fry migration at McClinton creek in the coming spring. - 15 -

A.L, Pritchard No, 11

Racial Studies of the Pink Salmon. —^————————mm I ii • »• i i •

It has long been maintained that the run of salmon to each stream is a separate entity in which the fish are characterized by definite anatomical or physiological characteristics. If the "parent stream theory" holds, i.e., if salmon return to spawn in the stream in which they were hatched, the existence of such definite and separ able races would be expected. In such a case there would be little chance of intermingling and cross fertilization which would tend to eliminate the differences. There are two common methods of investigating the occurrence of races. The first of these - marking - is direct. Numbers of young fish leaving the rivers are marked by the removal of certain fins. The location of the recoveries of the adult fish should indicate whether the group remains compact and returns to the same stream. Studies of this sort applied to the pink salmon at McClinton creek, Masset inlet, (see another report herein) have shown that wandering may occur, but it is apparently not widespread. The second method of studying population separation might be termed indirect. In this, counts or measurements of anatomical char acters are made in each area. These are then compared to determine whether there is any significant variation. If significant differences are discovered, one can be sure that the groups are separate. If the counts and measurements are the same, one cannot conclude that the two groups intermingle. It is possible that they may both have grown in sim ilar environment without mixing. This method has been applied with success to the clupeoids, viz., pilchards and herring. Recently in the case of the sockeye salmon indications have been given that the practice may work for the salmon. In this case the numbers of the gill rakers and the pyloric caecae appeared the best indicators. In order to test this method, which would evidently be cheaper than marking, certain collections of gill rakers and caecae were taken from pink salmon which had spawned and died in rivers on the Queen Char lotte islands. These are as follows:

McClinton creek, Masset inlet 250-300 Deenan river, mm £q Detlamen river, "" 60- 70 Mammon river, " " 100 Yakoun river, " " 100 Palent creek, Cumshewa inlet 8l

As soon as possible this material will be worked over. It is hoped that the results will indicate whether the new method can be used to support the results of the marking experiments. If the results are borne out, it is possible that it might be more economical to substitute this except in special cases where absolute returns are required. - 16 -

A.L, Pritchard No. 12

The Age of Chum Salmon in

Within recent years the importance of the chum salmon in the salmon fisheries of British Columbia has increased greatly. The species is now much sought both for canning and for the fresh market. It is obvious, therefore, that as complete information as possible on the life history of the species should be procured and made available both to the industry and those charged with its conservation. Furthermore, one of the most important phases of such information is that having to do with the age compositions of the runs. On this is based the whole idea of expectancy, which, though vague at the present, is the only basis on which programmes of productions may be planned. Studies of the age of chum salmon thus far in British Columbia have been sporadic and disconnected in nature. These, with possibly two * exceptions, have had as their goal the demonstration of the fact that scale patterns in the species will accurately reveal the age, and that growth increments in the scales from year to year may be used to cal culate the actual increase in length of the fish during those periods. In 1912 McMurrich interpreted a very small sample of scales. Almost at the same time J.A. Milne in England determined ages from a few scales which had been forwarded to him. Dr. Fraser in 1916 and 1917 collected a large amount of material in southern British Columbia. In one paper published in the Transactions of the Royal Society of Canada for 1919, he presents the age groupings and certain information on the relative growth of 1,300 individuals taken in the Qualicum district, and 700 captured around Nanaimo. In a second paper published in Contributions t0 Canadian Biology and Fisheries for 1918-1920, he submits data for a further collection, viz.- Qualicum - 506, Nanaimo - 379, and Chemainus - 139, In 1928 the present writer examined the scales of 67 chum salmon taken off the northern coast of the Queen Charlotte islands, and 33 from Qualicum. From the tagging programmes in Johnstone straits in 1929, 30 readings were made, and from those in the same area in 1930, 861. In 1933, 1934, and 1935, Dr. Hart in endeavouring to obtain per tinent information on the salmon runs to , worked over 255, 126, and 201 scales respectively. It is apparent that quite large numbers of chum salmon have been examined. Unfortunately, however, with the exception of the anal yses of the age of the tagged salmon which appear in Bulletin No. 31 of the Fisheries Research Board of Canada, the results have either not been published or appear in journals which are not easily procured by the industry and the conservation authorities. As the first step in the presentation of pertinent age data, the work which has been done should be summarized and made easily available. j If the locations of the collections made up to the present are examined, it will be immediately evident that certain areas which are now considered amongst the heaviest producers of the species do not appear. One example is the east coast of Moresby island, the southern member of the Queen Charlotte group. Certainly, therefore, the age - 17 -

A.X. Pritchard No. 12

survey should be extended to include this and others. A start has already been made with the collection of 332 scales at Pacofi cannery during the present season. These will be examined and every possible opportunity will be taken to make similar collections in future. It should perhaps be pointed out, that, while the study of such scale material may produce information of great value, the actual collec tion can be so arranged that it will entail little cost.

A.L. Pritchard and W.M. Cameron No. 13

Mortalities During Incubation of Chum Salmon Eggs in Stewart and Bush Creeks, near Ladysmith. and in the , , in 1939 and 1940

During January and February, 1940, in conjunction with certain laboratory experiments in which the investigators were attempting to im prove the procedure of redd sampling, the opportunity was taken to examine the mortality occurring during the incubation of chum salmon eggs in cer tain streams on Vancouver island close to Nanaimo, Nine redds were dug in Stewart and Bush creeks, near Ladysmith, and three in the lower portion of the Covaichan river near Duncan. Tho general findings are here briefly presented, Because of freshet conditions in the streams sampling could n©t be conducted during the period best suited for determining the efficiency of fertilization of the eggs - i.e. a few days after spawning was completed. The low mortality discovered later in some redds suggests, however, that the percentage fertility compared favourably with that previously reported for the pink and the sockeye salmon, viz.- approximately 98 per cent. The digging was carried out at the time when the eggs were in the last stages of incubation or the fry had just hatched. Mortalities up to that juncture are reported in the table below together with a brief des cription of the conditions which existed in each redd. As has been pointed out in an accompanying report, decomposition of dead eggs will weight the recovery figures obtained to an extent depen dent upon the proportion of heavily fungussed eggs or egg fragments, in the sample. To date no strictly accurate method has been devised for deter mining the relative recovery of these eggs from a particular nest. By com parison it would appear that the dead eggs and fragments in the chum sal mon samples should have been recovered at least as efficiently as were the coho salmon eggs in the experimental plantings referred to in the summary report mentioned above. The efficiency was here recorded as sixty per cent. To illustrate the effect which this decomposition might have on the results, a second column has been set up in the table for the Ladysmith eggs, presum ing that sixty per cent, only of the heavily fungussed eggs were recovered. - 18 -

A.L. Pritchard and W,M. Cameron No. 13

Redd Description Unweighted Weighted Mortality Mortality (Recovery of decom- (fc) posing eggs - 60$>)

Stewart and Bush creeks

1. Gravel fairly clean - medium size 7.6 9.4 2, Gravel fairly clean - medium size 11,1 17.0 3. Water shallow - silting moderate 42,0 49.0 4. Gravel fine - redd exposed at 85.2 87.5 normal water height 5. Gravel very clean - no plant debris £.6 3.0 or silt - strong current 6. Current moderate - redd below log 97.6 98.0 jam - much fine gravel, silt and debris 7. Heavily silted - fine sand 100,0 100.0 8. Large rocks - moderate to strong 2.0 2,0 current - no sand 9. Above fast riffle - gravel fine 4,4 7*2 fcut very loo*e - Alevins close to surface

Cowichan river

,1, Gravel medium ** redd on exposed 2,1 bar 2, Dry channel - water in redd bottom - 65.0 fair amount of silt and sand 5. Dry channel - good seepage - little 1.3 sand

The mortality sustained in the redds sampled varied widely. Apparently the difference was correlated to some extent with the relative amounts of silt and debris deposited on or among the eggs. Most favourable conditions for survival were associated with good circulation of water. - 19

W.A. Clemens jj0- 14

Contributions to the Life History of the Sockeye Salmon

The outstanding features of the sockeye salmon runs to Rivers inlet, Skeena river and Nass river in 1939 were as follows:

Rivers inlet ~ Data on 1,485 individuals (a) The 42 age group predominated constituting 67 per centj the 52 group was represented by 32 per cent and the 5* group by 1 per cent. This distribution was not an unusual one, (b) The average lengths and weights of the fish of both classes were much below the averages of the past years, particularly those of the four- year-old fish, (c) The percentages of the males and females was 50:50, but the number of males exceeded that of the females in the four-year-old age group and the females were greatly in the majority in the five-year-old age group.

Skeena river - Data on 1,290 individuals (a) The percentage distribution of the age groups was as follows: 4g»s - 50 per cent; 52*s ~ 35 per cent? 5jfs - 11 per centj 6**s - 4 per cent. This percentage of occurrence is approximately tho same as the average of the past years of record. ib) The average lengths in all the year classes were considerably above the averages of the past twenty-seven years of record while the average weights were decidedly below those of the past twenty-five years of record, (c) The males constituted a percentage of 38 and the females 62, The representation of the males was the lowest in twenty-five years of record,

Nass river - Date on 1,451 individuals (a) The percentage distribution of the age-groups was as follows: 42fs - 14 per cent; 52»s ~ 13 per centj 5,»s « 66 per centj the 6,»s ~ 7 per cent. The predominance of the 53 age-class was quite usual and almost identical with the average occurrence of the past twenty- seven years, namely, 67 per cent. (*>) The average lengths tended to be somewhat above those of past years of record while the weights were much below average. (c) The percentages of males and females were 48 and 52 respectively and practically identical with the averages of the past twenty-five years. - 20 -

A.L. Pritchard No^ ^

The Age of Spring Salmon in the Commercial Catches in British Columbia.

The 2,834 samples of scales which formed the basis for the pre sent study were collected during the years from 1927 to I930 in connec tion with the spring salmon tagging programmes being carried on in that period by the Fisheries Research Board of Canada. In certain of the general deductions which have been made, the writer has referred to pre vious analyses carried out on I926 and I927 material by Dr. C. McC. Mot- tley. Apparently most young spring 3almon migrate to sea as fry but some may stay in fresh water for at least a year. The former are refer red to as ocean type fish, and the latter as stream type. While there is a variation in the percentages of these types from area to area and from year to year as indicated by the scales, the average, reached on the basis of a large number of. fish, shows that about 78 per cent left the rivers as fry, while the remainder, approximately 22 per cent, remained until the yearling stage. Among the spring salmon caught off the coast of British Colum bia six age groups are definitely represented, namely, fish in their second, third, fourth, fifths sixth, and seventh years. A few samples indicate that some fish may even reach their eighth or ninth years. Those in their third, fourth, fifth, and sixth make up slightly over ninety-six per cent of the present collection. The ages for the "offshore" areas, such .as the west and north coasts of the Queen Charlotte islands, differ from those in the "inshore" districts such as off Stephens, Banks, and Porcher islands, Milbank and Fitzhugh sounds, and the north-east coast of Vancouver island. In the former there is a predominance of the older age groups, fish in their fifth and sixth years with a few in their seventh. The average age ranges from 5.0 to 5.6. In the latter, individuals in their third, fourth, and fifth years make up the major portion, the average age ranging from 3.7 to 4.4, The average age of the spring salmon in the commercial catches appears to become greater as the area of sampling moves from south to north as indicated by the following data: off Ucluelet ~ 3,8, off Kyuquot and Quatsino - 4.4, off the west coast of the Queen Charlotte islands - 5.5, and off the north coast of the same group ~ 5.1. Again spring salmon tagged along the coast and later recovered in the present a similar trend in the averages as follows: off Ucluelet - 4.2, off Kyuquot and Quatsino - 4.4, off the west coast of the Queen Charlotte islands - 5.3, and off the north coast of the same islands ~ 5.2. An explanation of the variation in age composition of the spring salmon catches from area to area has been devised after considering all available evidence from general observations, tagging, marking, and scale reading. The young spring salmon after leaving, their natal rivers, move up the coast on a relatively leisurely and slow migration which has been termed a "dispersal" of "feeding" movement. At some time in their subse quent life history they turn and return to the spawning streams. This "inward" or "spawning" migration is accomplished in a relatively short period. - 21 ~

A#Ii. Pritchard Ho, 15

Because spring salmon choose the larger rivers in which to spawn, the population along the coast is considered to consist of com paratively large groups which will dominate the fishery in the areas in which they are travelling although this domination may be masked some what by contributions from smaller coastwise tributaries. It also appears that the young spring salmon from the most southerly rivers do not travel so far north on their "feeding** migration before they return to spawn as do those from the more northeirly streams. It should be possible, therefore, to divide the coasteil area into sectors in each of which spring salmon from a certain a*iver will be dominant. Because of the extent and course of the migrations, these sectors will be relatively long and narrow and extend in a south-easterly to north-westerly direc tion. The boundary between one sector or sphere of dominance and the next will not be sharp since, as the salmoh from the southerly streams decrease in numbers, those from the more northerly will increase. Sampling of a population in any given area will present a crass-section of the immature fish migrating: away from the natal stream and of mature fish returning to spawn* The youngest spring salmon will be in the majority near the mouth of the stream which they have just left while the oldest will be preponderant et thS limit of migration. There will be a gradual increase in age between the first and the last extremesi If the contribution of a more northerly river enters, the occurrence will be indicated by the appearance of relatively large numbers of younger spring salmon. Complete detail of the findings and the discussion has been set down in a paper entitled* *»S-fcudies on the age bf the bono salmon COncor- hynchus kisutch) and of the spring salmon (OnOortiyllc^ tschawytseha) in British Columbia". This is how in press in the1' Transactions of the Royal Society of Canada for 1940, Section V.

A,L. Pritchard and A.L, Tester No. 16

Studies of the Food of Spring Salmon in British Columbia during 1940.

During 1939, 333 spring salmon stomachs were cellected in var ious areas in British Columbia. The results of the analyses have been summarized in the summary reports for that year and in Progress Report No. 42 of the Pacific Biological Station and the Pacific Fisheries Experimental Station. In 1940 an effort was made to increase the collectors and the range of collection, and thus raise the numbers of stomachs for examination. The following table presents a comparison of the effort in the two years: - 22 -

A.L, Pritchard and A.L. Tester No. 16

Area of Collection Number of Collectors 1939 1940

West coast of Vancouver island 2 5 and Central district 1 4 Prinoe Rupert and Northern Queen Charlotte is. area 6 7 Cowiohan bay ** 1 Campbell river - 1

To date the following stomachs have been received:

West coast of Vancouver island Strait of Georgia and Central district Prince Rupert and Northern Queen Charlotte is., area Cowlchen bay Campbell river

Since there are six collectors from the northern areas to still make returns, it appears fairly certain that the number will reach the objective of double that of 1939. In addition sampling was started ear lier in the season, continued later, and embraced more localities. The analyses should thus present a more reliable picture of the general food of the springs taken in the commercial catches throughout the season. Only 124 stomachs have been examined up to the present. These were all collected on the west coast of Vancouver island from Kyuquot to Borkley sound in the months of March, April, May, and June. It is strik ing how similar the contents are to those taken in the same locality in 1939* Sand lance was the predominant food at Kyuquot but occasional stomachs showed red feed. On April 16 and 17 fish were taken which had been eating young small pilchards which had appeared in great numbers off the this year for the first time. In the Berkley sound area sand lance were still found but there was a greater proportion of herring. The occurrence of young pilchards was more frequent. Ancho vies were also present. No conclusions can be drawn until the remainder of the stomachs are inspected. In the meantime every effort is being made to obtain material in as widely distributed areas as possible throughout the winter season.

j - 23 -

Ferris Neave No. 17

Returns of Marked Spring Salmon, Ccwichan River.

Approximately 50,000 spring salmon of brood years 1937 and 1938 were released in the Cowichan river in 1938 and 1939, It was hoped that some recoveries of these might be made this year. Unfortunately, owing to flood conditions, handling of fish at Skutz falls was practically elim inated. Consequently, examination of spring salmon was limited to those obtained by anglers in Cowichan bay and to random catches in the river. No marked fish were recovered from these sources. Since the majority of the mature fish returning to the Cowichan are fourth year fish, some might be expected to show up in 1941. Since most of the spring salmon returning to this river are fish which have gone to sea in their first spring, the marked fish, which were released in late summer and fall, will have spent one less summer in the sea than these. They may therefore be expected to be of smaller average size. Whether these factors influence later life history or affect the incidence in com mercial catches is not known. - 24 -

Ferris Neave No. 18

Returns of Marked Coho Salmon. Cowichan River

Brood Year 1957

Twenty-six thousand fingerlings raised at the hatchery and marked by the removal of adipose and left ventral fins were released in jfche upper Cowichan river in the fall of 1938 and early spring of 193p, Fifty-nine returns of fish marked in this manner (in addition to three of doubtful validity) have been received from salt water, most of these being from Cowichan bay and the Strait of Georgia as far north as Cape Mudge. Unfortunately fish marked in the same way were also re leased in the Lewis river, a tributary of the Columbia, Some doubt must therefore apply to individual records, though it seems unlikely that Columbia river fish would influence the general results tabulated below.

Locality Date No. of returns

Cowichan ba.y Aug. 1939 to Oct. 1940 35 April & August, 1940 2 Cape Mudge August, 1940 3 Lasqueti Is., & June to August, 1940 10 vicinity Dodd narrows August, 1940 l Maple bay September, 1940 1 Puget sound July & August,1940 2 Sooke October, 1940 1 Cape Flattery July & August, 1940 4

Twenty-two of the fish reported were caught in commercial operations, the remainder by anglers. Evidence of a considerable return of marked fish to the upper Cowichan river is provided by the following figures for cohos examined at certain points.

Locality Total no. of No. of marked cohos examined fish Skutz falls 165 7 Cowichan lake pool 99 35 Oliver creek 127 93

Of these localities a complete count is only possible at Oliver creek. The figures given for this stream to date probably represent only a small proportion of the total run.

Brood Year 1938

Fingerlings raised at the hatchery were released in Cowichan lake pool in Ootobor 1939, as follows: - 25 -

Ferris Neave No. 18

Lot 1, Offspring of third year males and females, 23,544, adipose and both ventrals removed. Lot, 2. Offspring of third year females and second year males, 8,567, dorsal and adipose removed.

Fish from these plantings have been recovered as follows:

Lot 1. Locality Date Number

Oliver & Beadnell April & May, 1940 54 creeks Cowichan bay * Aug, & Oct., 1940 3 Cowichan lake pool Oct., 1940 2 (sea run) Oliver creek Nov., 1940 1 ( " » )

60 Lot 2, Oliver & Beadnell April & May, 1940 12 Creeks Oliver creek Nov., 1940 2 (sea run) 14~

Ferris Heave No. 19

Production of Coho Fry in Two Small Streams

In continuation of previous investigations into the life his tory of the coho in the Cowichan river system, upstream and downstream migrants were counted on two small tributaries entering the Cowichan river about half a mile below Cowichan lake. Figures obtained for the spawning migration of 1939-40 were as Zallswsi

Oliver creek Beadnell

Total no, of cohos 841 582 No. of females 336 204 Dead, spent & stripped females 72 32 Females remaining 2$4 172 Eggs presumed available for depo sition, at 2,520 per female 665,280 433,440 Screens and traps for the collection of downstream migrants were maintained from March 13 to October 18. During this period the fol lowing figures were obtained for coho fry and fingerlings of the year: - 26 -

Ferris Neave No. 19

Number removed from traps Oliver creek Beadnell creek

Number removed from traps 78,789 174,056 Percentage of presumed deposition 11.84 40.01

The majority of the young fish entered the traps in April. The lower portions of Beadnell creek were dry, except for isolated small pools, from the end of June until October. A small number of young fish remained in Oliver creek when counting was discontinued. During their descent of the streams the young cohos were accom panied by yearlings of the same species and by trout. The following table shows the number of these potential predators obtained from the fry traps:

Coho Brown trout Cutthroat • yearlings Oliver creek 1,948 28 985 Beadnell creek 1,865 26 231 Cutthroat trout, which from stomach analyses appear to be the greatest individual consumers of young cohos in these streams, are much more numerous in Oliver creek than in Beadnell creek. It seems likely that this condition may account, at least in part, for the marked differences in percentage mortality which appears to have occurred in the two streams. All living fry from the traps were marked before liberation downstream, a different mark being used for each creek. Many of these fish soon dispersed widely. Marked fish were recovered in May from var ious localities in the Cowichan river from .he outlet of the lake to within two or three miles of salt water. The marked yearlings recovered from the small streams in April and May, had evidently ascended these waters previous to the installation of the screens on March 13. When captured they were proceeding downstream with other yearlings and fry. The tendency to move up small tributaries in winter or early spring is apparently found also in unmarked cohos, These have been found at the end of March in small ditches which are en tirely dry during most of the year and which are not ascended by adult fish. - 27 -

Ferris Neave No. 20

Salmon Angling in Cowichan Bay

As in 1939, a record was kept of angling activities in Cowi chan bay during the main fishing season, the purpose being to obtain information concerning the composition and relative size of the salmon run to the Cowichan river. An observer was stationed at the bay from August 1 to October 31 to make daily records of the number of boats and lines in use, length of time spent in fishing and size and species of fish caught. Records obtained may be summarized as follows:

Number of boats 3,203 " " lines 5,991 " " line-hours 21,236 Number of fish caught: Spring salmon large - 1,631 11 " "Jacks" - 356 Coho salmon large - 1,960 " " grilse - 423 Total fish caught - 4,370 Number of line-hours per fish - 4.8

Since in 1939 records were kept only from August 23 to October 31, a comparison for this period is presented.

1939 1940

Number of boats 2,803 2,464 " " lines 5,278 4,629 Number of line-hours 19,006 16,707 Number of fish caught: \,}- Spring salmon large - 553 1,121 " " "Jacks" - 750 292 Coho salmon large - 3,410 1,899 " " grilse - 1,069 160 Total fish caught - 5,782 3,472 Number of line-hours per fish - 3.3 4.8

The number of fish caught in 1940 was much smaller than in 1939 in each category except "large springs", of which approximately twice as many were taken as in the preceding season. The average weight of "large springs" was approximately the same in both years (17.5 pounds in 1939, 17.4 pounds in 1940). For purposes of the present classification fish of about 5 pounds and less are grouped as "jacks". The latter constituted only 17.9 per cent, of the spring salmon catch of 1940, as compared with - 28 -

Ferris Neave No. 20 56 per cent, in the previous season. Spring salmon fishing was very poor from September 15 to September 27, during which period a consider able mortality of this species occurred in the bay. Coho fishing was definitely much poorer than in 1939, both with respect to the number of fish caught and in terms of the effort required to obtain fish.

Ferris Neave No. 21

Trout Population of Cowichan Lake Pool

Cowichan lake pool is the most productive short stretch of Cowichan river as regards the number of trout taken annually by anglers. Fishing here is of a different type from that practised elsewhere on the river, being carried on almost entirely from small boats. Marked seasonal variations in the trout population take place. The following table has been compiled from records of gill-nettings dur ing the past three years. The figures given are limited to fish over nine inches in length, i.e. of a size commonly caught by anglers. This corjcesponds roughly to an age of two years or above. Season Number Percentage composition of catch of fisn Rainbow Cutthroat Brown trout

Jan.-Feb. 88 79.5 15.9 4,6 Mar.-Apr. 34 61.8 14.7 23.5 May-June 80 55 15 30 July-Aug. 142 19 13,4 67.6 Sept.-Oct. 77 70.1 11.7 18.2 Nov.-Dec. 29 86.2 13.8 0 The rainbow trout is the dominant game fish from September to May. The abundance of this species declines during late spring and early summer, when fish migrate to the lake. This decline corres ponds with a marked increase in the proportion of brown trout present in the pool. About the middle of September rainbows reappear in con siderable numbers. Brown trout are scarce in early winter, when many individuals have left the pool to spawn in the small tributaries, + * **. J* iS evident ttlat the °rown trout is fulfilling to some ex tent the hope that it would show a stronger inclination than the rainbow to remain in the river during the summer months. The tendency of many fcrown trout to remain in this general locality is further indicated W */ returns from a small planting of three hundred and three yearlings made in the pool in April, 1939. Thirty-six, or 11,88 per cent, have been recovered to date, all within a quarter of a mile of the point of release. Excellent growth is shown by these fish. Thirty-three had reached a length of eleven inches or more and individuals up to three and a half pounds in weight were, among those recovered. - 29 n

Ferris Neave No. 21 The incidence of brown trout in anglers' catches is not pro portional to the percentage occurrence of the species in gill-nets, being always much lower. Percentages of the three species of trout, based on anglers' catches in the pool in certain months, are shown below. Little fishing takes place during the summer when the proportion of brown trout is highest and little success has attended such effort as is made. It would seem that the feeding habits of this trout and possibly the me thods of fishing employed, afford the species relatively high protection from anglers.

Season Number Percentage composition of catch of fish Rainbow Cutthroat Brown trout Mar.-Apr. 332 87.35 11.75 0.9 May-June 62 79.03 16.13 4.84 Sept,-Oct. 80 97.5 1.25 1.25

Ferris Neave No. 22

Steelhead and Related Forms -a u ,, \seneral understanding of the relation between the sea-going and wholly fresh water forms of Salmo eairdneri is prerequisite for the adoption of a logical plan of fish management. U\ T^-iJ^f f011^1?6 typ6S °f fish are Present in the Cowichan river,- lyivwifknown lias steelheads.? M*? SP!n(?(b) Kamloopspart of theirtrout,livesraisedin attnethes*ahatcherya*<* a*e fromcommon egge collected in the interior of British Columbia. All those introduced f™if ll ?eTi6t? the Board'S invea^gations have been marked distinct- (c) irJ^f ? I" tnthe Cowicha11 **™r system are not yet known. spawnings.™?j£i Sincec S W£1C?Kamloopsd° n0ttroutg° weret0 sealiberatedand whichinto™ the systemresultedinfromformernatural leill8a+l «%m? b6.an adraixture of stock, but it seems certain that«non- Tnnlnfh in'this+ i f+existedcategory areinmthegeneralCowichanmuchbeforemore coarsely,these Planti^sscaled «•than Kam-^L. angers?^a^.j^fd'r1.^d6Signated as "rainb™rVtagetrout", StthethetermhatChe^universallyFor usedconvenience,by local in th« rJ?idefable numbers of marked Kamloops trout have been liberated these n^vb^TTooltTr *7V*? dU3?ing the paSt three years and re^»s ^m 1940?o!£ .approximately7 . ?6d55,000f0r durlngsteelheadtne forthcomingfingerlingsfishingand sfsooseason.fingeringsDuringfrom eZTZTZelZtT11are being released, each''I**.type"rainbows"being marked^ distinctively,*-n «i«d at theGentryobject beineand to determine the migratory or non-migratory tendencies of these forms. § upperunner part^t fl?^Tof the f^ichant™™river,,rainbows"has seemeda^ towinter-runreveal certainsteelheadsaveragein thedif- - 30 -

Ferris Neave No. 22 ferences in scale counts, suggesting that the tendency towards one or other mode of life might be hereditary or else established at an early age. For the purpose of the following comparisons, a "rainbow trout" is defined as a fish at the end of its third summer, or older, which shows no indication of sea life. Since temperature conditions during early development have been shown to have effect on the number of scales, the differences observed in "wild" fish might conceivably be due to this factor. Eggs from each type of fish were therefore taken and raised under the same temperature conditions in the hatchery. Samples from these lots are included in the following table:

No. of scales • "Wild" fish Hatchery fish on lat. line Steelhead Rainbow Steelhead Rainbow

117 .» 1

118 - 4 «» _

119 - 9 0* ..

120 - 12 - 4 121 2 8 1 6 122 3 4 0 5 123 5 3 3 5 124 7 3 4 4 125 3 - 3 1 126 2 m. 4 127 1 m. 1 _

128 0 .. 2 «» 129 1 OT 3 ^

130 a. — 1 -

£4 fish 44 fis;h 22 fish 25 fish Av, 1^4.25 Av. 120.39 Av. 125.65 Av. 122.08 - 31 -

G.C. Carl No. 23

Cowichan Lake Hatchery During the past season the hatchery has J>een operated as usual to supply the fish necessary for the marking and tagging programme of the Cowichan investigation and as a means of studying hatchery methods. Construction and maintenance of fish trans. For the collection of parent fish, traps were maintained in Oliver and Beadnell creeks near, the hatchery and in Bear creek flowing from . Fry traps were installed on the downstream side of the weirs in the first-named streams for the enum eration and marking of coho. The fences and traps formerly operated in Holmes and Crossland creeks, Somenos lake, were removed in i\fey. During the low-water period in September a more permanent trap was installed at the upper end of the Skutz Falls fish-way for the purposes of enumerating the salmon run, recording marked fish and collecting spring salmon for eggs. EggL.?olleotion. The nuntow of egg5 collected during tne past year are as follows:-

Spring 17,000 Steelhead 84,000 Coho 67,000 Rainbow 4,000 Brown 2,700 Cutthroat 2,600 In addition to the above, the hatchery received 100,000 eggs of Kamloops trout in June through the courtesy of the Provincial Game Department. Liberations and fish on hmrifl. The number of fish reared, marked and re leased as fry, fingerlings or yearlings, in the Cowichan river system in 1940 is as follows:-

Species Stage Brood Number Mark

Steelhead Fingerlings 1939 34,800 Adipose, right ventral Coho Fry 1939 50,000 None Cutthroat Fingerlings 1939 S,117 Adipose, right ventral Kamloops Fingerlings 1S39 25,550 Adipose, left' ventral Brown Fingerlings 1939 1,632 Adipose, right ventral Spring Fingerlings 1939 9,393 Dorsal, left ventral Kamloops Yearlings 1939 1,900 Adipose, left ventral (some tagged) In addition to the fish handled by the hatchery a total of 244,000 coho fry were trapped, marked and released in Oliver and Beadnell streams dur ing March and April. In May, 8,000 Kamloops fingerlings received from the Provincial Game Department were marked by removal of the adipose and left ventral fins and were released. - 32 -

G.C. Carl No. 23

Fish diseases. Two disease organisms occurred among the Kamloops fry during the summer causing heavy losses about the middle of August, The first organism recognized was Costia, a ciliated protozoan present in large numbers on the surface of the backs and sides of the infected fish. A portion of the fry, given a formalin treatment (1 part to 4,000 for 1 hour) as prescribed by Fish (Progi Jish-Culturistj Jan. 1940), suffered heavy mortality which led to the discovery of a second disease-producing organism, Octomitus. in the intestinal tract. The fish were then given the only treatment that appears to be known, 2.7 gms. of calomel added to each 3 lbs. of meat used in the diet, (Wright, Prog. Fish-Culturist, #£4, Nov. 1936) but with little noticeable improvement. The two diseases finally ran their courses with a total loss of over 90 per cent. Asphyxia in ponds. Bni August 24, a heavy loss was found to have occurred in one pond of Steelhead fingerlings but no cause was at first evident. However, since subsequent tests showed that oxygen values in the pond in question had fallen to 3,0 cc. per litre at the upper end (59 per cent, saturated) and 2.3 cc./l at the lower end (45 per cent, sat.) suffocation was thought to be the cause of death. Further tests showed that whereas ample dissolved oxygen was present in the reservoir (90 per cent, sat.) and in the filter tank (100 per cent, sat.) the supply was much reduced after leaving the filter (60 - 65 per cent. sat.). Cleaning the filter resulted in a rise of oxygen saturation to 85 per cent, in the filtered water which suggested that dissolved oxygen was being removed from the system possibly by decomposing material in the filter sand. Cleaning the filter at more frequent intervals (alternate days instead of twice a week) eliminated further trouble from asphyxia. Diet and food costs. During the current year the fry in the hatchery were fed finely ground fresh beef heart, while fingerlings in the outdoor rearing ponds received a mixture of several ingredients. The finger- ling diet was similar to that used in former seasons, the approximate composition being:- fresh meat (liver or liver and heart combined) 36 parts, dry ingredients (equal parts of skimmed milk powder, pilchard meal and cottonseed meal) 36 parts, stock mineral mixture 1 part and water 27 parts. The cost per pound of the mixture, including delivery charges, varied from 12 to 16 cents depending on the cost of the fresh meat and on the proportions used. The approximate cost in food per fish for three lots released this season was as follows:- Steelhead trout (34,800 fish; age 50 weeks) §.0085 Kamloops trout (25,550 fish; age 47-49 weeks) ^.0087 Spring salmon ( 9,388 fish; age 29 weeks) &.0042 It will be noted that in the case of the trout which were retained over winter the food costs were about twice that of the salmon which were held only during the spring and summer. In addition to the above a group of 1900 of the larger sized Kamloops were retained for about 16 months and received raw beef during the last 4 months in addition to the regular mixture giving a total food cost per fish of ^.0963. Closed water system. The system of re-circulating the water through a fil ter to the rearing ponds was employed again during the period of July to "• 33 **

G.C. Carl . No< 2S

October whenever the flow from the gravity supply was found to be insuf ficient. Before the period of operation a new type of piston ring was installed and a high grade lubricating oil was adopted as recommended by the distributors, but considerable trouble was experienced during the season with faulty performance and frequent breakdowns. The total time was approximately 69|- days at a total cost of fc>107.£0 (including alter ations and repairs) which gives an average daily running cost of $1.54 as compared with ^1.01 for 1939.

G.C, Carl No. 24

Limnology of Cowichan Lake

The limnological phase of the Cowichan investigation was con tinued this season by taking the usual series of observations on lake and river conditions. It was found that a minimum surface temperature of 6.5°C» was attained about Feb. 29, 1940, while a maximum lake sur face temperature of 21,4°C. obtained about July 22. The bottom tem perature (35 metres) rose from 6.3°C. to &»1°C. during the season and a thermocline formed between the 10 and 20 metre levels about the end of June. Quantitative plankton trap samples in vertical series have been obtained along with oxygen and pH determinations. The mooring buoy placed at the sampling station last year has been maintained except for two short periods when it was carried away by log booms. During a period of low water in August a hydrometric station was established at the hatchery boathouse under the supervision of the Dominion Hydrographic Branch for the purpose of measuring the rates of flow and total annual run-off of the Cowichan river.

G,C. Carl No. £5

Incubator Experiments

Some of the experiments carried out in 1939 were repeated this season in a further attempt to determine if any advantages result from incubating fish eggs at temperatures higher than those of the hatchery water. A second incubator was constructed using a drip of water from an unheated source in order to compare results with those obtained with the thermostatically controlled incubator. The experimental material consisted of two lots of coho eggs, each lot being divided into 3 portions, one of each lot being placed in the heated incubator, one in the unheated in cubator and one in the hatchery trough as a control. At time of hatching the eggs from both incubators were transferred to baskets in a hatchery trough except those of one lot (Bx) which were placed in heated water aerated by a small compression pump. The results are tabulated as follows;- - 34 -

No. 3& G.C. Carl Si

Lot Yjhere Number Incubation Temperature Loss in number placed of eggs period range eggs

Heated inc. 844 7 weeks 7 - 15°C. 27,8% 556 it ti » i» 100.0$ BlJb Av. 12°C. A Unheated »' 912 10 ?i 5 - 11°C. 6.756 B2 (1 1! 656 h it n it 0.9>5 Av. 8°C. A3 Hatchery 981 12 it 4 - 8°C. 2.5% < M 757 n « «t it 1.6% Av. 6°C.

In general it will be seen that the incubator eggs hatched sooner than the corresponding lots held under hatchery conditions and that the incubation period appears to be inversely related to the average temperature. A part of the heavy loss experienced by the eggs in the heated incubator was due to a rapidly growing fungus but the greater part occurred after the transfer to water and possibly re sulted from a too sudden change in temperature or from asphyxia. The experiments were not carried into the free-swimming stage.

G.C. Carl No. fk

Observations on Certain Competitors and Predators of Trout and Salmon in the Cowichan River System.

Coincident with the major problems which involve the propa gation and life histories of the salmon in trout in the Cowichan river system, the opportunity has been presented to make certain observations on other animals which were suspected of being either predators upon or competitors of the species under investigation. These data are sub mitted herein in the hope that the whole story may eventually be built up. Sticklebacks. During the year the stomachs of 178 sticklebacks were ex amined. The list of food items in order of importance included: Copepods, aquatic insects, cladocera, ostracods, algae, amphipods, mites, rotifers, protozoa, pisidium, and terrestrial insects. There is no in dication of difference between the food of yearlings, young or adults. Late in the season certain organisms sueh as ostracods appear more often. This may be purely the result of greater availability. In any event it has been demonstrated that the sticklebacks do compete with the trout and salmon for food. Kingfishers* During August and September kingfishers appear to become more numerous when the liberations of hatchery fish are being made into the river. Twelve birds were trapped near the hatchery. The food items and the number of occurrences were: sticklebacks - 12, salmonids - 7, insects - 4, pebble - 1, and plant tissue - 1. The insects included one ~ 35 -

G.C. Carl No. m caddis larva, one dyticcid beetle, and one nymph and one adult water bug (Lethocerus sp,) Lampreys. Lampreys collected in Holmes creek in 1939 have been iden tified by Dr. C.W. Creaser of Wayne University, Detroit, Michigan, as Lampetra planeri Bloch. This is the first known record of the occurrence of the species on Vancouver island in British Columbia. Another species, Entosphenus tridentatus. is also found in considerable numbers in the Cowichan system. It appears to be parasitic for a time of fresh water fishes after metamorphosis and before it goes to sea. Due to adverse weather conditions it was impossible to make quantitative observations on the run of Lampetra to Holmes creek. Qualitative observations in April and May, however, failed to reveal nests or breeding lampreys. It is therefore presumed that fewer individuals were present than in the previous year. - 36 -

J.Li Hart and A.L. Tester No, 27

The Insertion and Recovery of Herring Tags.

The programme of herring tagging and the recovery of herring tags designed to give information on the movements of herring and the existence and stability of local populations is being continued; The methods employed are the same as those used in previous years: herring are tagged by the use of numbered internal tags made of magnetic metal and recovery is accomplished by means of magnets or by the use of in duction detectors at the herring reduction plants. The merits and short comings of these methods have been dealt with in previous reports. Be sides expense, unreliability of operation has been the chief drawback of the induction detector type of apparatus. Accordingly, considerable time and money have been spent in attempts to work out a method of de tection which would be thoroughly reliable. Unfortunately it must be reported that so far no principle on which detection may be based has been found more practical than the one at present employed. It is hoped, however, that changes in the detail of construction made in the original sets will lead to minor improvements in operation. The results of the recovery operations conducted during the fishery of 1939-40 have been presented rather fully in a report published by the British Columbia Fisheries Department. It is pointed out there that if the herring fishing grounds are divided into four major areas, yiz., the east coast of Vancouver island including Deepwater bay, the west coast of Vancouver island, the central coast line, and the northern coast line, 99, 98, 99, and 93 per cent, respectively, of the tags re covered in them have been put out in the same area. This finding sup ports the view already put forward that the runs to the major areas tend to form distinct units. Evidence was obtained which showed that during the past season fish taken in the area had been on the east-coast-of- Vancouver-island spawning grounds during the previous spawning season. In view of the particular interest attaching at the present time to the herring situation on the west coast of Vancouver island, a tabular summary of the results from that area has been prepared in such a way as to indicate the extent of intermixture of tagged herring between its subdivisions.

i •do ra cd 13 «H +3 • cd -P t* to • fd © © © >» • t* to u u u +> © «d W cd Qi © • Q d w o u > cd o -h rd -p o H OTO^j^CQ .rt pi M i» -P a* -P to .h d in • m-d o ? m -p o cd

Total: 52 - 79 5 400 84 620 - 37 -

J.L. Hart and A.L. Tester No. 27

This table shows that interpretations are offered for 536 returns. Of these, 134 were tagged in and recovered in west coast areas, 397 were tagged in and recovered in areas other than west coast, 3 were tagged in west coast areas and recovered from fish caught in other major areas, and 2 tags used outside of the west coast region were recovered on west coast fishing grounds. In considering the 134 tags put out and recov ered on the west coast it will be noted that 42 out of 43 of the tags used in were recovered there, 72 out of 77 of the tags put out in the spawning areas around Nootka and Kyuquot sounds were recov ered in the same areas, and all 4 of the Quatsino tags recovered were returned from the Quatsino fishing area. Of the 10 tags put out in the -Sydney inlet area, 5 were returned from Barkley sound and 3 from Nootka-Kyuquot fish. In considering this result consideration should be given to two facts which together are significant: the pro portion of tags recovered to those put out was small in this area; no commercial catches of herring were made in the area. There was some indication of tags used in Nootka sound being returned from there and of tags used in being recovered from Kyuquot sound fish, but the evidence is not clear-cut. It is apparent that mixing between the subdivisions of the west coast of Vancouver island is limited. The investigation is being continued. During the spring of 1940, 28,303 herring were tagged. Notable changes from the tagging of previous years were the comparative absence of tagging on the fishing grounds and the extension of tagging operations into newly exploited regions such as the Queen Charlotte islands (400 tags) and the Queen Charlotte sound region where 3,796 tags were put out. It is doubtful if with the present financial resources of the investigation it will be possible to include a thorough investigation of the herring runs around the Queen Charlotte islands. Returns from some of the taggings done during the spring of 1940 are already being made. They agree with those of the previous years in indicating the near-segregation of major fishing grounds and the ten dency for some stability among the respective herring populations on the west coast of Vancouver island. It is worthy of note that the Queen Char lotte sound area so far appears distinct from all others. However, evi dence on this point is not yet sufficient for even tentative conclusions. The herring tagging programme was assisted by the loan of boats for tagging operations by the Nootka Packing Co. (1937) Ltd., and by Nel son Bros. Fisheries Limited, and has enjoyed the cooperation of operators and fishermen in other ways. It may be pointed out that one method of canning which is being widely used during the current year will leave most of the tags in the canned fish. The Board is apprised of this situation owing to the possi bility of operators becoming involved in damage suits based upon the find ing of the metal tags in the canned product. - 38 -

A.L. Tester. No. 28

Catch Statistics of the Herring Fishery during 1939-40.

The use of Pilot House Record Books for the collection of daily catch statistics of the herring fishery was continued during the 1939-4-0 season. The object of this investigation is to determine as accurately as possible the quantity of herring removed from each fishing area and the fishing effort expended. From these, the availability of herring is calculated as the average catch per seine per day*s active fishing. Availability figures are of prime importance in tracing fluctuations and trends in abundance. The total catch in 1939-40, over 150,000 tons, was the largest in the history of the fishery and considerably surpassed the 1938-39 re cords of approximately 107,000 tons. Of the total catch, 32f. was taken on old-established fishing grounds on the east and west coasts of Van couver island, 13f» on new grounds adjacent to the north-east coast of the island, 37f. along the central coastline, 13f. along the northern coast line, and 3fo on the south-east coast of the Queen Charlotte islands. As usual, most of the catch was reduced to oil and meal, but there was a noteworthy increase in the quantity canned (413,659 cases). On the south-east coast of Vancouver island herring were some what less abundant than during the previous season and individual sets were smaller on the average. The quota was extended from 25,000 to 35»000 tons, but the latter figure was not reached. In addition to double seiners, a large number of single seiners fished in this area, chiefly in and Nanoose bay. Availability figures for double seining from 1933-34 to 1939-40 follow: 76.4, 94.8, 101.0, 87.3, 144.0, 98.2, and 83.6 tons. During late January and the first half of February particularly successful fishing was encountered at Deepwater bay and Okisollo channel at the north end of the strait of Georgia. In this, and in the Alert bay area, the season was extended from Feb. 5 to Feb. 15. Tagging work sug gests that these fish belonged in part at least to the same population which supplies the south-east coast fishery. Although fishing was slightly more successful than in the previous year on the west coast of Vancouver island, it was still very poor and individual quotas were not attained in any area. Stormy weather during the fall hindered fishing to some extent. Availability figures for the last five seasons in the area from Cape Beale to Cape Cook are as follows: 67.9, 50.2, 31.6, 12.7, and 17.5 tons. Prospects for the 1940-41 season are possibly brighter than those for 1939-40, but there are still no indications of particularly successful fishing. During the fall, several seiners prospected the waters adjacent to Queen Charlotte sound (Alert bay area) for the first time, with a con siderable degree of success. The more promising: individual grounds were Belleisle sound and Clio channel, A new reduction plant was built at Alert bay and was in operation by. the first week in January. Herring were somewhat less available in the area during the latter half of the season. Fall fishing in the central area (Klemtu passage, Cousins inlet, Rivers inlet, etc.) was attempted, but with poor results. During - 39 -

A.L. Tester No. 28

early January herring were again encountered at Kwakshua passage and the adjacent waters of FitzHugh sound and good catches were made. The availability progressively decreased throughout January and February, however. On the average the availability (78 tons) was considerably less than that of last year (over 111 tons) when spectacular fishing was encountered. From January to March, Laredo inlet proved to be most produc tive. More boats were fishing and more fish were caught in this inlet than on any other individual ground. The availability was high, aver aging 83 tons during the ten weeks of fishing, and the abundance ap peared to be maintained until the date of closure. Prince Rupert harbour again yielded very successful fishing, although the period of great abundance was confined to one month (Feb ruary) . During March, boats operating from Rupert plants prospected much of the northern area and fished as far north as Wales passage, at the Alaskan boundary. Availability figures for Rupert harbour from 1937-38 to 1939-40 are as follows: 94.6, 55.3, and 93.5 tons. Summer fishing was attempted on the east and west coasts of the Queen Charlotte islands, but the availability was very low (11 tons). Inskip channel on the west coast yielded most of the catches, but the fish were scattered and not abundant. From December until the middle of March, seiners operated on the east coast with considerable success. The availability of herring (42 tons) was comparable with that usually en countered in single seining on old-established fishing grounds in the vicinity of Vancouver island. During the 1939-40 season, daily records of 75$ of the total catch were submitted by seine boat captains through the medium of Pilot House Record Books. In order to secure more complete returns the cooper ation of the Department of Fisheries has been secured in distributing the books and in warning those concerned regarding penalties which might be imposed for failure to submit the required statistics.

A.L. Tester No. 29

Herring Spawning Conditions in 1940.

With the cooperation of the Dominion Department of Fisheries, herring spawning reports compiled by fishery inspectors were again col lected in the 1940 season. The object of this investigation is primarily to obtain information on the intensity of spawning in each area in order that fluctuations or trends in spawning intensity may be detected. Infor mation is also given on the dates of individual spawnings, mortality of the eggs, weather conditions, etc. The data for 1940 have been summarized in tabular form, giving for each area the dates of the first and last individual spawnings, the average intensity as reported, the calculated acreage of spawning ground, and a calculated intensity index. Intensities are reported in five cate gories or combinations of categories ranging from very light (VL) to very - 40 -

A,L. Tester No. 29 heavy (VH). Tho acreage has been calculated roughly, allowing 12 acres of spawn to the mile of spawning beach (i.e., an average width of 100 feet to the spawning ground), Tho intensity index is the sum of the acreage times the intensity expressed numerically from 1 (VL) to 5 (VH) for individual spawnings. The data for 1940, with comparative inten sity figures for 1939, follow;

Calculated Calculated Intensity Inten acres Index Area Time Spread sity 1940 1939 1940 1939

Victoria Mar. 5-Apr.l5 M 6 5 15 17*~ 9 Cowichan Feb. 4-Mar, 5 L 24 15 49 35 Ladysmith Feb.21-Apr.12 H 238 101 1018 454 Nanaimo Mar. 3-Apr. 4 VH 126 144 548 501 Comox Feb.l5-Apr. 4 MH 90 246 324 1035 Quathiaski Apr. 2-Apr. 3 H 36 36 135 117M*m>mmm, J Pender har. Mar. 3-Apr. 9 H 104 65 398 213 Strait of "——* Georgia 623 613 2487 2372 Barkley Feb.14-Mar.29 L 82 147 223 577 Clayoquot Feb. 2-Apr.ll M 111 18 365 36 Nootka Fob. 2-Apr. 8 M 70 10 214 24 Kyuquot Feb.l3-Mar,30 L 14 14 39 36 Quatsino Feb.24-Mar,10 L 38 43 94 112 West coast 1 ••••- Vancouver is. 315 232 935 785 Alert b. Feb.17-Apr.10 H 445 235 I698 844 Bella Bella Bella Coola Mar. 4-Apr.l6 M 319 510 1109 1008 Rivers in.

Butedale Mar.22-Apr. 2+ MH 246? 336 858? 1488

Grenville- Principe Mar.15- M 13? 16? 30? 45? Skeena Mar.20-Apr. 6 H 168 115 624 411

North Q.C.I, Jun.25- H 48 48 192 168 South Q.C.I, Mar,29-May 5 H 336 18? 1332 45? - 41 -

A*L* Tester '. No. 29

The quantitative estimates of acreage and the intensity indices miist be regarded as approximatej but in most cases are considered to be more reliable for yearly comparisons than the opinions of individuals. The data for areas north of Vancouver island are probably incomplete, particularly in the Butedale and Grenviile-Principe areas, owing to the difficulties of making adequate surveys. In most general districts spawning appeared to be somewhat heavier in 1940 than in 1939* One probable exception is the Butedale area, but this may be due to incomplete information. In the strait of Georgia, the heaviest spawning in 1940 was in tho Ladysmith area, whereas in 1939 it was in the Comox area. On the west coast of Vancouver island spawning in the Clayoquot and Nootka areas was considerably heavier, but in the Barkley, Kyuquot, and Quatsino areas it was lighter than in 1939. The considerably greater spawning acreage of the strait of Georgia as compared with the west coast is worthy of note. It supports the suggestion that tho potential herring productivity of the former is greater. The high intensity index for the Alert bay area in 1940 may have been due in part to a more complete survey. At heavy spawning was reported at intervals along a 20-mile shoreline. Heavy spawning was reported on the southern Queen Charlotte islands, an area which was investigated in detail for the first time. Attention is called to the fact that no February spawnings are reported in areas north of Vancouver island. The relatively late April spawnings at Quathiaski and the usually late June spawnings at Masset inlet, north Q.C.I., are also worthy of note. In the strait of Georgia most reports agree that mortality of the eggs was lower than usual because of favourable weather conditions and a reduced number of feeding birds. However, ducks caused a high mor tality in the Cowichan area. As usual, the feeding of birds caused high mortality on the west coast of Vancouver island and also on some spawning grounds in the central area. February storms were also destructive on the west coast of Vancouver island. Frost during the spawning period in the central and probably the northern area did not constitute the usual hazard in 1940*

A.L. Tester. No> '^ Time of Herring Spawning along the West Coast of Vancouver Islands In District No. 3 (Vancouver island and adjacent mainland waters) herring seining closes each year on Feb. 5, a date which is fixed by the Department of Fisheries to protect spawning herring and to prevent the disturbance of herring spawning grounds. In District No. 2 (north of Vancouver island) where spawning takes place somewhat later, there is no fixed date of closure, each area being closed at the discretion of the local inspector. During 1940, certain members of the industry suggested - 42 -

A.L. Tester No- 30 that the Department of Fisheries abolish the definite closing date for herring fishing in the Quatsino aroa on the west coast of Vancouver is land, contending ttfat the spawning period in this area was more closely allied to that of District No. 2 than to District No. 3. In response to a roquest from the Department, information regarding the dates of spawning along the west coast of Vancouver island over a ten-year period was tabulated and the following comments were submitted: 1, There is considerable variation in each locality from year to year regarding time of spawning. During a spell of good weather in March spawning may take place simultaneously in all areas from Barkley to Quatsino. 2, The records indicate that, in general, the main spawning period is earlier in Barkley sound (about Feb. 25 to Mar, 15) than at (about Mar. 5 to Mar. 25). 3. At Barkley sound, in five years over a ten-year period, spawnings have been reported before Feb. 5» At Clayoquot and Kyuquot, over the same period, spawnings have been reported before Feb. 5 ia one or two years. In all areas south-west of Cape Cook there are several reports of spawnings early in February (after Feb, 5). 4. At Quatsino sound proper (Browning inlet, Winter harbour, Koprino, etc.) the earliest spawning recorded is Feb. 20. In all other years of a ten-year period spawnings have been in March* 5. In the Brooks bay and Klaskish region, which is also in cluded in the Quatsino area, spawning takes place earlier than at Quat sino sound proper. There is one report of spawning on Feb. 12 and several reports for Mar. 1 and 2. This year (1940) herring spawned on Feb, 24, or possibly earlier, 6, In areas to the north-west of Vancouver island there are no records of spawning in February (although small isolated spawnings may possibly have taken place)*, 7» Although spawning in the Quatsino area tends to be some what later than in other areas along the west coast of Vancouver island, in view of the several reports of February spawnings (particularly at Klaskish), it seems to be more closely allied to District No. 3 than District No* 2 in respect to time of herring spawning.

A.L. Tester No. 31

The Sampling of Herring in Southern British Columbia,

The sampling of herring is carried on with the object of tracing fluctuations in abundance which may be caused by either year class vari ation or the action of the fishery, and of determining the extent of intermingling of the runs by a study of variation in racial characters. The continued expansion of the fishery and the broadened scope of the s.cientific. investigation has made necessary an even greater col laboration than in previous years between the various herring investigators. N

• 43 -

A,L. Tester No. 31

All have contributed to the collection of data in the field. As a re sult of this combined effort, 90 samples totalling 9,485 fish were ex amined from southern waters alone during 1939-40, These were distri buted as follows:

Area Season No. • Sampled by:

Queen Charlotte Fishing 14 Boughton, Tester, sound Spawning 3 Boughton, Tester,

West coast of Fishing 29 Hart, McHugh, Quickenden, Tester. Vancouver is. Spawning 7 Tester, McHugh,

East coast of Fishing 29 Boughton, Tester. Vancouver is. Spawning 8 Tester, McHugh.

The samples taken, during the spawning season were from schools which were caught and tagged. These should enable comparisons to be made be tween the runs present during the fishing season and those actually spawning in each area. They may also help to relate the results of racial studies with those of tagging in determining the extent of inter mingling- of the fish. The time-consuming task of age estimation has been completed up to and including the season 1939-40. The summarizing and analysis of the mass of data collected since the start of the second period of inves tigation (1935-36) is well under way, but must await the inclusion of the 1940-41 data before it can be completed. If this information is to be available for a general survey of the herring situation which is to be made in 1941t additional clerical assistance will be necessary in the near future. The results to date are embodied in three separate reports, each dealing v/ith one of the three major fishing areas of southern Bri- ^ tish Columbia. These follow this general report. /

A.L. Tester No. 32

The Herring of Queen Charlotte Sound and Adjacent Waters.

In 1939-40, the herring fishery exploited the waters of the Queen Charlotte sound area for the first time. Good catches were made in Belleisle sound off Kingcomo inlet, at Knight inlet near Minstrel is land, and in Clio channel near Bones bay, as well as scattered catches from other localities. Samples from each major fishing ground were salted at the Alert bay plant and shipped to the Station for examination. The herring from these new fishing grounds were for the most part very small, averaging (after correction for salt shrinkage) 171.7 mm. for Belleisle sound, 158.7 mm. for Knight inlet, and 183.4 for Clio chan nel, as compared with 187.3 mm. for the south-east coast of Vancouver is land. One sample from Retreat pass, however, had an exceptionally large average length of 218.3 mm. - 44 -

A.L, Tester No.. 32 ^ectod to f^ <* «" *i*. it was rather tage of fish of tho otder age ~l t ** excoPtio^^ large percen- followingcomparison:tabulation, toge^e/with tha^te^8"1??8tnat for the south-eastm coastSh0Wn forin the

Locality u IXI Iy No. TO VII VIII IX+ Samples Belleisle sd, 22*8 12.3 4 1 21 * */:*,, —«f -•„. SS'* SiSS iiI:,0 sa,H B»sH J;|« £>:: ?-1y ? &.e. coast 26.3«••? PP./53 1 10.7lni 6.0£ n o2,2« 0.9 0.1 - 15 ^^.ZiW.£^^*«^ °* H... the dominance of II's, V-f andt^ \ r^111 Sh0Wing ^ outstanding aofremarkablyadvanced apeUtf'pL^C^n.. (7 to xtt) »^t i?i£*pf? Retreatrt Sami)lespass sample«»*«*«*fish III IT v VI VII samples aSTK."- g-j gg ;«:« jg-| »j ««., , R—CliO Ch.— 15*3'-(C 811t/7/: o £:!,«„ .. KJ-OU.O S:i- 204-!„ .I 3 S.e. coast 164.4 1,0., 201., 20?.0 215.7 219.5 15

fishofth^tSis particularlyZTstrikingZ'^Zlrate^erowth£L°? ^^ * the Be^oisloless th- «**sound smaller, than fish of ego III o'n !£?A?80 *? at Bel^isle sound are It might be noted that tLLU*?~°aS^ .TSt °f V**°°™>r island, were quite distinctive In that thnv J f *heSe 8l°»-S«"l»« fish

32=:of the south-eastSSsr.r£Scoast rfTS„ ilP^? ?* average counts- than"*-8"those toStrikingworrfntdifferencesinthe coLlusion^hataL ™»Boll^^*• ^^""* ^°wthmUS' rate,to8e*oris sufficientwith the population which is distort L ^f10 S0Und herrinS constitute a 5= S=Ss£^*tritt3!i-"-JS2 - 45 -

A.L, Tester .No. 32

The presence of a few individual fish of relatively large size with a rather vague scale pattern suggests a small degree of mixing of fast-growing fish such as those of the strait of'Georgia with the pecul iar slow-growing type of the Queen Charlotte sound area, particularly in the case of Knight inlet and Clio channel. This slight admixture may account for the apparent difference in growth rates between the fish of the latter two localities and Belleisle sound.

A.L. Tester No. 33

The Herring of the West Coast of Vancouver Island.

Efforts are being made to coordinate the sampling data gathered over the past five years on the west coast of Vancouver island. When the data for the current season are available they will be combined with those of previous years preparatory to a general survey of the west coast situation in 1941, It should be pointed out that the season 1940- 41 terminates a five-year experimental period of regulation of the fishery by the quota system. Racial data are being used to supplement the results of tag ging to determine tho extent of intermingling of the various west coast runs. An analysis of vertebral counts which is at present underway shows certain differences which are of, or approaching, statistical significance between west coast areas. Such differences clearly indi cate that several year-classes of Quatsino sound herring could not have intermingled freely with those of Barkley, Clayoquot, Nootka, and Kyuquot sounds. This supports the conclusion drawn from previous racial work and from tagging work regarding the essential isolation of the Quatsino sound run. In the remaining west coast localities, significant differences are found between Barkley and Kyuquot sounds (1931 and 1938 yr. cl.), between Barkley and Nootka sounds (1934 yr.'cl.j, and between Nootka and Kyuquot sounds (1938 yr, cl.). These indicate that the fish of certain year classes, at least, have not intermingled freely. On the other hand, at both Nootka and Kyuquot sounds there are cases of sig nificant or nearly significant differences between fish of the same year class, but of different ages. For example, at Nootka sound in 1938-39, II*s of the 1937 year class had an average count of 52.082, whereas in 1939-40, Ill's of the same year class had an average count of 51.794. The significant difference between these two values indicates that fish of the same year class were.not drawn from the same population in two successive years at Nootka sound. This, then, is evidence of some degree of mixing. The problem remains to determine how extensive is the mixing between areas other than Quatsino sound. It is doubtful if vertebral comparisons alone can add materially to the solution of this problem. Attempts are being made at the present time to group samples according to similarities in length and age composition with the hope of obtaining some indication of the affinities of aberrant samples. This involves - 46 -

A.L. Tester No. 35 the assumption that individual schools do not merge. If this attempt fails, the one remaining possibility is to try to group individuals, and the only character which suggests itself as suitable for such a classification is scale pattern. If a study of scale pattern does not lead to positive results, further contributions to the study of the de gree of intermingling will depend entirely on the results of tagging work. Studies of age composition of west coast runs show considerable variation between areas in the same season. For example, in 1939-40, Ill's predominated in the majority of the Barkley sound samples (59.4f.), whereas II's predominated in all of the Nootka sound samples (66.9$). In addition, significant differences may occasionally occur between the age compositions of individual samples in the same locality and season. This variation,"together with the doubt surrounding the extent of inter mingling between areas, renders difficult the task of making a general survey of annual changes in age composition. At Quatsino sound, which may be considered to constitute a local run, year classes 1931, 1933 1935, and 1937 have been relatively abundant and have "followed through" in a uniform manner from 1935-36 to 1939-40. In the latter year, however, a sample taken from the spawning grounds showed a predominance of II»s (1938 yr. cl.). If the data for the remaining west coast areas are weighted to the catch and combined, the following data result:

__ „ ^ IV, & Availa- Year II III iv V VI VII VIII IX older bility 1936-37 6.9 46,5 35.9 6.7 3.3 0.2 o.l 0.1 46.3 50.2 193J-38 4.8 60,1 16,5 14.5 2.4 1.2 0.3 0.1 35.0 31.6 1938-39 43.8 34.5 17.9 2.0 1,5 0.2 -- 21.6 12 7 1939-40 36.6 43.0 12.8 6.2 0.8 0.4 0.1 - 20.1 17.5 Coincident with the falling-off in the availability of fish to the fishermen there has been a relative decrease in the abundance of fish of age IV and older, and in 1938-39 and 1939-40 a spectacular increase in the relative abundance of fish of age II. The latter were not only pre cociously mature, but were also large for their age. In both 1938-39 and 1939-40, II's have been present in considerable abundance in areas other than the west coast, suggesting an unusual premature entry of II»s for some unknown reason. However, on the west coast such a premature addition would have been expected to increase the availability of herring in 1938-39 if the older classes were present in average abundance. It is difficult to escape the conclusion that there has been a great de cline in the actual abundance of older fish and that the 1937 and 1938 year classes, which were present as II«s in 1938-39 and 1939-40 were relatively but not actually of great abundance. A basis for this con clusion is contained in the following figures which are considered to be proportional to the actual numerical abundance of fish. In the calcula tions, it has been assumed that availability (catch per seine per day) is proportional to abundance, an assumption which has some justification in the case of the west coast: - 47 -

No. 33 A.L. Tester • Proportional Numbers

. IV & Year II III IV V VI VII VIII IX over

1936-37 379 2552 1970 368 181 11 6 5 2541 1937-38 166 2080 571 502 83 42 10 3 1211

1938-39 •679 505 177 31 23 3 - - 234

1939-40 750 882 262 127 16 8 2 - 415

The decline in the abundance of fish of age IV and over, and the unex pectedly small numbers (c.f. percentages above) of II's in 1938-39 and 1939-40 are worthy of note. It was mentioned above that the II*s which have predominated in the catches for the past two years have been unusually large. This is illustrated by the following average lengths for fish of each age in Barkley sound:

Year II III XV V VI

1936-36 164.3 188.4 199.6 207.3 213.8 1936-37 152.8 187,9 W.7 207,3 213.6 1937-38 158.3 184.9 . 201.5 207.6 215.I 1938-39 171.9 187.9 201.8 207,8 216.7 1939-40 170,9 19^0 203.2 210.7 221J.

Not only are the IIfs in 1938-39 and 1939-40 th4e largest on record, but the Ill's in 1939-40 are also outstandingly large, as might be ex« pected if the same unusual group of fish returned to Barkley sound in the second year. The increased length of other age groups in 1939-40 is also worthy of note, suggesting the addition of older fish whioh were subject to similar growth-promoting conditions as the II's. Some difficulty has been experienced in diagnosing the scales of some of the fish of this unusual group. This might be expected, for it is characteristic of fast-growing fish to have poorly defined annuli on the scales. However, at present, a fair degree of confidence is felt in the reliability of the interpretations. The study will be continued with the inclusion of the 1940- 41 data and a detailed account of the results and conclusions will be prepared as soon as possible.

A.L. Tester No. 34

Herring of the East Coast of Vancouver Island.

It has already been demonstrated by tagging work that the herring which supply the fishery on the south-east coast of Vancouver island are derived from numerous spawning grounds along both shores of the strait of Georgia. It is doubtful if much additional information can be obtained from a detailed study of racial characters. However, cer tain features of the runs to particular localities are worthy of note* - 48 -

A,L. Tester No. 34

In both 1938-39 and 1939-40 the writer's attention has been called to the large size of the fish taken at Nanoose bay. The average age composition of tho samples shows that fish of an advanced age wore relatively more abundant than in the Swanson-Trincomali-Stuart channel area to the south. In the two years, fish of age IV and over formed 5lf. and 40$ at Nanoose as compared with 30f. and 20f. in the southern area. Racial characters showed no differences which could be regarded as clear- cut evidence of a local run, although the difference in vertebral count of one year class (1936) was on the borderline of statistical signifi cance and tho older fish were exceptionally large for their age (see below), Why the schools of older fish should have tended to congregate here is unknown. Possibly it is an accidental occurrence, for in past years (1934-35 and previous) Nanoose bay fish were of average size and in both 1938-39 and 1939-40 occasional samples of similar size were taken in both localities. There has been some speculation as to the origin of the her ring fished for the first time at Deepwater bay at the north ond of the strait of Georgia.- It is fairly certain that the fish of this area contained tags inserted in the strait of Georgia, leading to the con clusion that they belonged, in part at least, to the population of that area,- However, differences in growth rate are shown by the follow ing data:

Area II III IV V VI VTI

Deepwater b. 158.5 179.9 197.0 206.1 212.2 215.4 Swanson ch., etc. 164.4 190.7 201.7 209.0 215.7 219.5 Nanoose b. 159.8 189,4 202.7 211.5 221.2 227.5

The average size of the fish at each age at Deepwater bay is less than that on the south-east coast of Vancouver island.. This is particularly the case with the Ill's which also showed a distinctive scale pattern. This suggests that the Deepwater bay run was a mixture of fish from different areas - the strait of Georgia, and possibly Queen Charlotte sound which has been shown elsewhere to have runs of slow-growing fish. The following table gives the age composition of herring of the Swanson-Trincomali-Stuart channel area over the past five years:

Year II Hi IV V VT VII VIII+ 1935-36 1.3 39.9 31.5 20.5 1.3 4.2 1.3 1936-37 5.3 55.0 23.4 10.8 3.9 1.0 0,8 1937-38 1.8 52.3 28.2 12.3 4.0 1.3 '1938-39 28.2 35.8 23.1 7-9 3.3 1.0 0,7 1939-40 26,3 53.7 10.7 6.0 2.2 0.9 0.1

It is apparent that on the south-east coast as well as well as on the west coast of Vancouver island, there.was an unusual relative abundance of II's in both 1938-39 and 1939-40. However, the percentages were - 49 -

A.L. Tester No. 34 lower on the east coast and the size of the fish was also smaller and more nearly normal. On both coasts the average length of the fish of each ago was greater in 1939-40 than in previous years. In view of the falling off in the availability of herring on the south-east coast from 1937-38 to 1939-40, it seems fairly certain that the presence of II's in 1938-39 and 1939-40 represents an unusually early addition of fish rather than being an indication of the presence of a particularly abundant year class. These precocious II's are simi lar to, but not identical with, those which occurred on the west coast. On both coasts the II's differed from those which occurred at Kwakshua passage in 1938-39 and Queen Charlotte sound in 1939-40 in being for the most part mature »ather than immature. The results for the current season (1940-41) may possibly help to explain the unusual changes in age composition which have taken place.

R.V. Boughton Ncu ,5 Length and Age Composition of the Herring Catch in Northern B.C. ^ Sampling of the commercial herring catch has been continued in the northern waters of B.C. during the past season. The purpose of the investigation is to trace natural fluctuations in abundance caused by variation in the strength of successive year classes. The data col lected during the past season shows some interesting results when com pared with the data of last year. A total of 37 samples was taken from various localities throughout the northern waters of the province during the past fishing season. * The areas exploited were the same as last year: first the central area, in which the major portion of the catch was taken from Kwakshua passage and Laredo inlet (smaller amounts were taken from Klemtu passage and Jackson pass, but no samples were examined from the latter place); second, the northern area, in which practically all the catch was w^L^Whales channelT and?P!LrtSwansonharbour>bay; withand, third,«* exceptionthe Queenof aCharlottesmall amountislands,from where fish were taken during the summer from Inskip channel and Scudder point, and during the winter from Skaat harbour, Sewell inlet, and Lagoon +•„>,-. distinct differences occurred in both length and age composi tion between samples from the three major areas and considerable varia- o^L°nwfQueen Charlotte* betweeAsamplesislands and infromthe centrala^erentarea.AshingIn averagegrounds lengtharound the samples may be arranged in the following order: Inskip channel, 217.5 mm.. Scudder point, 202.7 mm.; Laredo inlet, 199.7 mm.; Skaat harbour, ' 196.5 mm.; Prince Rupert harbour, 194.3 mm.; Kwakshua passage, 193.6 mm • Sewell inlet-lagoon bay, 188.4 mm.; and Klemtu passage! 143?7 mm. Th7 great variation between localities is evident. - 50 -

R.V. Boughton No. 35

Distinct differences in age composition are also apparent. These are illustrated by the following figures giving average and dom inant age groups (latter in brackets): Inskip channel, 7.3(V); Scudder point, 5.7(vl); Laredo inlet, 5.5(VI); Prince Rupert harbour, 5,1(VI); Skaat harbour, 5.0(111); Kwakshua passage, 4,4(111): Sewell inlet-Lagoon bay, 4.2(111); and Klemtu passage, 2,5(11). It is impossible to compare the length and age compositions of the samples taken from the Queen Charlotte islands with those of last year, because fish were caught in different localities. In all othor areas adjacent to the mainland from which fish were examined there has been an increase in both average length and age. This increase is shown by the following figures (average age in brackets): Prince Rupert harbour, 194.3, 192.0 (5.1, 4.4); Klemtu passage, 143.7, 130.1 (2,5, 2.3); Laredo inlet, 199.7, 196.8 ip'.p", 4.7); and Kwakshua passage, 193.6, 180.0 (4.4, 3.6). Two possibilities exist to account for the increase in average length and age during the past season. There may have been either a decrease in the normal number of young fish entering the fishery, or an increase in the normal number of older fish present. Studies in age composition appear to favour the former possibility. These studies sug gest that in 1938-39 the number of herring belonging to the younger year classes which entered the fishery was above normal, certainly in the case of the 1937 year class at Kwakshua and possibly in the case of the 1936 year class from Rupert harbour. Fish belonging to the same year classes apparently failed to return in normal numbers in 1939-40. This failure to return may have been caused by unfavourable oceanographic conditions, by migration to other areas, or by fishermen capturing a large proportion in 1938-39. Little is known regarding the first two possibilities, but the last was undoubtedly a contributing factor. In any case, the failure of the younger year classes to return in 1939-40 would naturally result in the fish increasing in average length and age. It is proposed that sampling of the commercial herring catch be continued in the northern area.

R.V. Boughton No. 36

Racial Studies on Herring North of Vancouver Island.

With a view to procuring evidence of the extent of inter mingling of herring runs in northern B.C., studies have been continued on racial characters. The runs which were sampled during the past sea son have been outlined in another report. Significant differences in average vertebral counts of year classes from different areas indicate that intermingling of herring is limited or absent between runs. The accompanying table shows the differ ences in average vertebral counts of year classes from different areas and the "statistical significance" of these differences. The table does not include comparisons which do not show significant differences. - 51 -

R.V, Boughton No. 36

Year Difference in Localities Comparod: Class Average Counts D/S.E.D

Rupert : Kwakshua 1935 .198 3.65 Rupert : Kwakshua 1936 .429 4.83 Rupert : Laredo 1935 .210 ?.8l Ruport : Laredo 1936 .223 3.09 Rupert : Sewell-Lagoon 1934 .420 3.21 Kwakshua : Sewell-Lagoon 1934 .498 3.34 Ruport : Klemtu 1937 .244 2.35 Laredo : Sewell-Lagoon 1934 .345 2.74 Laredo : Klemtu 1937 .301 2.91 Sewell-Lagoon : Skaat 1934 .516 3.41 Sewell-Lagoon : Scudder 1934 .523 2.93

By an examination of the table it will be noted that differ ences in vertebral counts indicate lack of extensive intermingling be tween the following runs: Prince Rupert harbour and Kwakshua passage (19351 1936 yr, els.); Prince Rupert harbour and Laredo inlet (1935, 1936 yr. els.); Prince Rupert harbour and Sewell inlet-Lagoon bay (Q.C.I.) (1934 yr, cl,); Kwakshua passage and Sewell inlet-Lagoon bay (1934 yr. cl.); Prince Rupert harbour and Klemtu passage (1937 yr. cl.); Laredo inlet and Sewall inlet-Lagoon bay (1934 yr. cl.); Laredo inlet and Klemtu passage (1937 yr. cl.); Sewell inlet-Lagoon bay and Skaat harbour (1934 yr. cl.); and Sewell inlet-Lagoon bay and Scudder point (1934 yr. cl.)* The differences noted above indicate that intermingling of herring was limited or absent between runs to the Queen Charlotte is lands, the northern and central areas. Those same conclusions were ar rived at last year on the basis of vertebral counts and ago composition. The table further indicates that local runs may exist within the central and Queen Charlotte island areas. Studies in age composition also suggest the isolation of cer tain of the runs sampled. The accompanying table gives the percentage age composition of the various runs sampled during the 1939-40 season.

Locality II III IV V VT VII VIII IX+ Av.

Kwakshua 1.4 37.2 8.6 35.0 11.5 5.0 1.0 0.2 5.0 Laredo 0.7 10.8 12.1 23.6 30.1 15.7 3.5 3.5 5.3 Rupert - 12.6 22.5 26.5 29.3 4,8 2,4 2.5 5.1 Klemtu 56.0 37.6 5.8 0.5 -- - - 2.5 Skaat - 31.6 6.9 26.2 17.0 7.4 4.1 6.3 5.4 Inskip - 1.1 3.3 22.2 15.6 17.8 5.6 34.5 7.3 Scudder - 8.3 13.5 21.9 34.5 13.5 2.1 6.2 5.7 Sewell- 6.0 31.6 12.1 19.1 14.9 7,4 4,6 4.1 4.7 Lagoon -52 -

R.V. Boughton No> 36 From the table it *is evident that tho runs adjacent to the mainland have distinct differences in their age composition. In the Queen Charlotte islands differences in age composition suggest the isolation of runs to Skaat harbour and Sewell inlet-Lagoon bay (• a winter fishery) from those to Inskip channel and Scudder point ( a summer fishery). That intermingling is limited or absent between the runs to. Scudder point and Inskip channel is suggested by the sample from Inskip containing a very large percentage of fish oldor than eight years of age and having an average age of 7,3 years. The age composi tion of the samples from Skaat harbour and Sewell inlet-Lagoon bay are somewhat similar except for the lack of two-year-old fish in the Skaat harbour run. It has been shown, however, that there is a signi ficant difference in vertebral counts between these two runs. Racial studies will be continued in northern B.C. The data will be analyzed in an effort to determine fluctuations in abundance and the extent of intermingling between runs.

R.V. Boughton Nc.# -^

Length and Weight Differences between Laredo and Prince Rupert Herrinj; — K

During the 1939-40 herring season it was apparent from casual observation that there was a distinct difference in the appearance of fish from the run to Laredo inlet and that to Prince Rupert harbour. The Laredo herring were "fatter" than those from Prince Rupert harbour. The difference observed was due, in all probability, to differences in the length-weight and age relationships. The follov/ing table shows the average lengths and weights for fish of various ages from Laredo inlet and Prince Rupert harbour.

Average Length Average Weight

Age Laredo in. Rupert har* Laredo in. Rupert har.

II 140.3 mm 28.9 _ JII 181*8 174.6 75.1 62.9 IV 195.8 189..2 97.2 83.8 ¥ 201,8 195.1 109.2 93.5 VI 210.1 201,2 125.2 104.3 VII 212,0 206.4 130.6 115.0 ¥111 217.0 206.3 140,7 118.5 IX 216.5 212.1 143.1 119.1 The figures clearly demonstrate that herring of a given age from Laredo inlet were longer and heavier than fish of the same age from Prince Rupert' harbour. For example, herring that were in their third -53 -

R.V, Boughton No. 37

year had an average difference of approximately 7.2 mm. in length and 20,9 em, in weight. By the time Laredo herring had reached their third year they were longer and .heavier than those from Prince Rupert harbour, In the case of length this advantage was maintained throughout life, whereas in the case of weight the advantage was not only maintained but was progressively increased with each succeeding year. The following table demonstrates very strikingly that fish of the same age and length from the Laredo run were heavier than those from Prince Rupert harbour (with one exception, starred),

Age IV Age V Age VI Age VII

•Lar Rup Lar Rup Lar Rup Lar Rup Length, edo ort edo ert edo ert edo ert

.. m ^ — ^. 173-177 60.6 71.8 «,

178-182 72,0 73.6* 8O.5 72,7 - 74.7 - —

183-187 81.2 78,9 85.1 79,4 • 81.5 - —

188-192 86.8 84.9 92.6 86.6 97.4 89.6 - 94.8 193-197 97.2 89.7 98.7 94.3 104.1 96.4 108.6 98,0 198-202 102;3 93.9 105.8 97.7 109.2 102.0 112.7 98.0 203-207 110.6 105.8 112.7 105.6 117.9 106,2 114.5 112.7 208-212 119.0 113.0 123.4 U6.5 124.1 114,3 125,0 120.9

213-217 - - 127,8 - 132.4 125.2 137.3 125.6

w 218-222 - 137.6 - 140.9 138.9 144,6 143.4

m 223-227 - <•> - 148.8 142.0 156,0 -

The average differences for the IV, V, VI, and VII-year-old herring are approximately 4,9, 6,6, 7,6, and 7.4 gm., respectively. It may be noted further that in each age group fish in the smaller length brackets from Prince Rupert harbour have no comparable group at Laredo inlet. The consistent difference in the length-weight and age relation ship suggests that the Laredo herring were in better "condition" than those from Prince Rupert harbour. It might be assumed that differences in condition reflect, to some extent, differences in oil or fat content. The records of one plant bear out this assumption, for the Laredo fish gave a higher oil yield. Thus the difference in appearance between the fish from Lar edo inlet and Rupert harbour was due, in all probability, to the length- weight and age relationships. These consistent and striking differences indicate that the two runs must have spent the early summer and fall, in 1939 at least, on separate feeding grounds, and therefore comprised distinct groups or units. The superior condition of the Laredo herring over the Rupert herring may have been caused by differences either in quantity or quality of the available food supply and probably other, en vironmental factors. Analyses of length-weight and age relationships may yield considerable information on the extent of intermingling of the runs, and consequently will be continued in the future. - 54 -

A.L, Tester No. 38

The Use of the Echo-Sounder in Locating Schools of Herring.

Some interest has been shown recently in the possible use of the echo-sounder in locating schools of herring in B.C. waters. This interest has been partly stimulated f»rom claims of success published in descriptive circulars by the manufacturers of this instrument. In response to a request from the industry, efforts were made in 1939 to contact fisheries scientists in Europe to obtain information on the practical usefulness of this device in locating fish in European waters. Letters of inquiry were dispatched to Qscar Sund, Norway; W.C. Hodgson, E. Ford, and F.M. Davis, England; G.P. Farran, Ireland; and Henry Wood, Scotland. Replies were received from all except the first. According to the replies, there is no knowledge of the echo- sounder having been put to practical use in locating schools of fish in tho waters adjacent to the British Isles with one exception - the herring fishing boat "Nobles", which was reported to have installed the equipment and to have used it to advantage. Dr. Wood reports that numerous unsuccessful trials have been made to locate herring with an osho-sounder on the research vessel of the Fishery Board for Scotland on drift net grounds both off the east and west coasts of Scotland during summer and in the Forth estuary during winter. He states that "the success of the device for locating shoals depends mainly on local conditions. Two essentials v/ould appear to be a considerable depth of water and some degree of layering of the fish in the immediate water layers". It seems probable that the two essential conditions for the successful use of the instrument in locating schools of fish, as out lined by Dr. Wood, are realized on some B.C. herring fishing grounds during the fall and winter. It is a question, however, whether the suc cess which might be encountered in its use would justify the expense entailed in purchasing and installing the equipment. If further infor mation is desired by the industry it might be possible to make arrange ments for experimental echo-soundings on herring fishing grounds with equipment already installed on government hydrographic survey ships.

R.V. Boughton and A.L. Tester No. 39

Herring Egg Count and Weights.

With the intensive exploitation of herring and the possibility of serious reduction in the stocks along the B.C. coast, a study of the number of eggs per female herring is of considerable importance. It has long been known that in other species there is a tendency for egg count to increase with increase in length and weight of the fish. A decline in length and age composition such as would be expected under conditions of intensive exploitation would therefore result in a de crease in the number of eggs deposited on the spawning grounds because of the loss of the larger and older fish with the greater egg produc tion. This decrease would be in addition to that caused by the reduc tion in the number of spawning fish. - 55 -

R.V, Boughton and A.L. Tester No. 39

During the 1939-40 season the ovaries of 101 females were collected from various localities north of Vancouver island. After preservation in 10$ formalin the ovaries were weighed and a small sample of each re<-weighed and countedi Thus the total number of eggs per fish was determined♦ The dsta from 26 determinations made in 1933 were also available for use and comparison.

Egg counts All data illustrated a marked tendency for the average number of eggs per fish to increase with length. The number was found to vary from an average of approximately 10,000 in fish of 170 mm, to 35»000 in fish of 235 mm. In terms of age, the number varied from an average of approximately 12,000 for Ill's to 30,000 for VTII's. It was found that the egg count for a given length varies in different localities. In 10 out of 11 length classes that were compared, Laredo herring had a higher egg count than Prince Rupert her ring. This difference, in all probability, is due to variation in "condition" of the fish. The following table shows the relationship between length, egg count, and condition:

Length Class Condition Egg Count Condition Egg Count

171-180 2,0 7,100 2.7 15,453 181-190 2,0 9,300 2.7 17,300 191-200 2.0 15,300 2.7 20,400 201-210 2.0 15,800 2.7 24,200 211-220 2.0 21,200 2.7 27,100 221-230 2.0 23,700 2,7 34,300

The table, containing only a portion of the available data, which has been smoothed by the method of least squares, shows (1) that for fish of the same condition egg count increases with length, and (2) that for fish of the same length egg count increases with condition. A three- dimensional graph illustrating this complex relationship has been con structed.

Egg weight Egg weights were determined at the same time as egg counts. The correlation of egg weight with other factors such as length, weight, egg count, etc., is difficult with the limited data on hand, as egg weight is determined largely by the degree of maturity. Seasonal change in egg weight is illustrated by the following example. At Laredo inlet, the egg weights of fish belonging to the 180 ram. class increased from 0.000432 gm. on Jan. 15 to 0.000825 gm. on Mar. 15. Similarly, the egg weight of fish belonging to the 215 mm. class increased from 0.000541 gm, to 0.001050 gni. Therefore, in less than two months the eggs almost doubled in weight. Correlation of egg weight with other factors is further com plicated by the fact that individuals within the same school and taken at the same time differ in their maturity. At Butler cove, for example, the eggs of fish of 170 mm. varied in weight from 0.000735 gm. to 0.000971 gm., depending on the degree of maturity. - 56 -

R.V. Boughton and A.L. Tester No. 39

The data further shows that in fish taken at. the same time, but from different areas, there is a tendency for a difference in'egg weight to occur. For example, the average egg weight of Prince Rupert herring was greater than that of Laredo herring for fish of approxi mately the same length in 9 out of 12 comparisons. As far as could be determined, the fish in the two areas were of approximately the same stage of maturity. This study will be continued in the future as opportunity permits.

R.V. Boughton No. 40

Fish Preservatives, with Special Reference to Herring.

During the past year an effort was made to procure a suitable preservative for herring. At the present time herring, when shipped, are usually dry-salted or put into brine. The salt apparently des troys the ligaments which connect the centra and consequently make it impossible to remove the vertebral column as a whole when boiled. To obtain the vertebral count it is therefore necessary to "slice" the fish and count the vertebrae with the column partly embedded in the flesh. This method is not entirely satisfactory. Preliminary tests were made on a number of solutions which appeared to have possibilities as a preservative. The following solu tions were used: tannic acid (5, 10, and 15$); potassium permanganate (5 and 10$); concentrated salt solutions with 5 and 30$ glycerine; and phenol (1, 3, and 5$). In all cases except those involving the 3 and 5$ phenol solutions there was almost complete putrefaction of the fish, caused either by fungus or bacterial grovrbh. In the case of the phenol solutions, however, the fish were in excellent condition. Externally the fish appeared just as well preserved as specimens placed in alcohol and the use of phenol as a preservative for museum specimens appears possible. Phenol costs considerably less than alcohol and does not evaporate as readily. In an attempt to find the cheapest preservative further tests were carried out. Twenty fish were placed in each of the following: crude phenol (5$); crude cresol (3$); and phenol (crystals U.S. P. 5$). The pure phenol solution, however, was the only one in which the speci mens did not putrefy. The fish in this solution were in excellent condition and the vertebral columns were removed in the usual manner. From tests recorded it was definitely established that the pure phenol solution was superior to all others tested in prohibiting putrefaction. A further test was then made to determine the amount of shrinkage in length and weight which occurred over a period of two weeks. The method of least squares was used to procure the relationship between the length and weight of the fish and tho decrease due to pre servation. The data showed a tendency for the amount of shrinkage to - 57 -

R.V. Boughton No. 40

increase with standard length and to decrease with weight. Shrinkage in length varied from 4,4 mm. for fish of 140 mm, to 4,6 mm, for fish of 250 mm. Loss of weight varied from 5.1 gm, for fish of 60 gm. to approximately 4.8 gm. for fish of 120 gm. For all practical purposes a mean shrinkage of 4.5 mm. could be added to the standard length and 5.0 gm, to the standard weight of preserved fish so as to make the re sults comparable with measurements from fresh samples. The possibility that the cumulative effects of the weak phenol solution over a period of time might produce a skin irritation still exists, but it is probable that this can be overcome by the ad-, dition of a neutralizing agent before handling the fish. Additional tests are being carried out to check tho results already obtained and to determine whether further shrinkage occurs.

J.L. McHugh No. 41

Vertebral Variation in Young Pacific Herring.

The study of abnormal vertebrae in young herring commenced in 1939 and discussed in summary report No, 31 for that year has been completed during the current year and the results prepared for publica tion. The types of complex vertebral elements observed were as follows:

Type 1. Two centra fused, with the completeness of fusion greater on one side of the body than on the other.

Type 2, Suture spiral in form, so that an extra vertebra appears to be present on one side of the body.

Type 3. A series of centra heavily calcified and rigidly fused to gether .

Type 4, Bilaterally symmetrical multiple centra.

Type 5. Duplications of spines in the posterior two centra.

Since complex vertebrae increase in frequency up to a certain point as the larvae increase in size, it is concluded that abnormalities are the result of fusion in the outer bony rings of the centrum. Skeletons containing abnormalities of type 5 give counts which are consistently lower than those for "normal" backbones. On the other hand, in vertebral columns containing abnormalities of types 1 to 4, if each element involved in a fusion is counted as one, the count is consistently higher than for "normal" skeletons. It appears, there fore, that abnormalities of type 5 should be considered as fusions of whole vertebrae, and that if tho count obtained from "abnormal" back bones is correct, the true vertebral number of a population lies some where between the values for "normal" and "abnormal" vertebral columns. - 58 -

*J.L. McHugh No. 41

Comparing the frequency of complex vertebrae in samples from Nanoose and Departure bays with that for samples from the southern areas combined, a significant difference is found to exist (t » 5,92; P *c 0,01), This is apparently related to water temperatures, the areas having a lower average temperature producing populations with a lower frequency of complex vertebrae. The frequency with which abnormal elements occur is great est in the region of the twentieth vertebra. Investigation of the movements of the fins during metamorphosis shows that no direct as sociation between fin movements and frequency of abnormality is evident. A comparison of the frequency of abnormality along the ver tebral column with the variability in length of the individual centra shows that a significant negative correlation exists (r » -0,55; P<0.01). The above data seem to point to differential growth as the causative agent in producing abnormal centra. Pressures set up by the action of growth-gradients in the development of the vertebral column and in the surrounding tissues provide a satisfactory explanation which is supported by the known facts concerning the structure and de velopment of the contra.

J*L. McHugh No, 42

Investigations Dealing with Vertebral Number in Young Herring.

With the exception of Departure bay, no sampling of young herring schools was carried out during 1940. Further progress was made, however, toward completing the examination of material from the collections of previous years and analysing the accumulated data. The samples examined are summarized in the following table:

Year No. Vert. $ Locality Class Fish Ncl. 0 °M Ab, Departure b. 1939 86 51.802 0.601 O.067 5.8 193 51.742 0.585 0,044 7.8 Maple b# 1939 342 51.905 0,698 0.039 6.7 Fulford. har* 1939 105 51.694 0.665 0.067 6.7 80 52.053 O.691 0.079 5.0 Bedwell har. 1939 201 51*835 0.976 0.105 6.5

Data involving the average vertebral number of a population do not take into account the change in form and function of succeed ing centra. By dividing the vertebral column into its various parts it is possible to provide additional information. Abdominal and caudal vertebrae are those having open and closed haemal arches respectively. -59 -

J,L. McHugh No. 42 Tester has shown that a significant difference exists between abdom inal counts of fish from Barkley sound and Saltspring island, although no difference exists in total count. In the young herring it was found that the abdominal and caudal regions of the backbone vary in almost equal proportions between localities in the strait of Georgia, and a similar change with vertebral count takes place within individ ual samples, "Tail" vertebrae are those having the haemal arches "cross- tied" to the centrum. The values for this segment of the backbone are more erratic and apparently do not vary in proportion to the total count. The following table gives average tail counts for several samples. The samples from Nanoose bay and Fulford harbour, although differing only slightly in total count, differ significantly in the number of tail vertebrae.

Number of Fish with tail count of; Locality Year 4 3 6 7 Mean

Fulford har. 1935 8 161 & ?• 5.199 Fulford har. 1937 25 184 79 5 5.206 Ladysmith har. 1937 9 178 108 4 5.358 Cowichan b. 1938 1 23 14 2 5.425 5.602 Nanoose b. 1934 - .31 44 3 Further evidence that the larger fish in a sample tend to have higher vertebral counts has been brought forth. By arranging each sample in order of increasing length, and dividing into two equal parts, the higher length group shows a higher average number of centra, as il lustrated in the following table:

Average Count of Length Groups Locality Year Small Large

Nanoose b. 1934 51.66 51.80 Fulford har. 1935 51.71 51.88 Fulford har. 1937 51.69 51.92 Ladysmith har. 1937 51.83 51.98 Presdy har. 1938 51.72 51.82 Fulford har. 1938 51.91 52.12 * 1939 51.90 52.01 BidwsU b. 1939 51.88 52.09 - 60 -

J.L. McHugh No. 43

Growth of Young Herring in Departure Bay, 1940.

Weekly samples of young herring were again taken in Depar ture bay by means of an electric lamp and a lift net. The larvae were apparently very numerous this year, but disappeared about a month after hatching and did not reappear in the bay until early fall. It may be significant that their disappearance coincided with the ar rival of large schools of anchovies and pilchards, which remained in the vicinity throughout the summer. At least two modes were present in the length frequency curves for each sample, but these were not nearly so distinct as those observed in the 1939 material. A study of the general growth trend, however, shows that the average increase in length was somewhat more rapid than for the 1939 year class. Water temperatures were slightly higher in Departure bay during the period of early development in 1940 than in 1939, and this may have been a contributing factor. The general trend of growth and water temperatures is illustrated in the following tabulation: Av, Length Smoothed water in mm, temp, °C 1939 1940 1939 , 1940 At hatching 8.5 8.0 9.2 9.9 1st. week 11.0 10.5 9.8 10.3 2nd, week 13.0 13,0 10.2 10.8 3rd. week 15,0 17.O 10.7 11.2 4th* week 17,5 20.5 11.2 11.7

J.L. McHugh No. 44

Experimental Tagging and Retention of Young Herring. Experiments conducted during the fall of 1939 and recorded in summary report No. 34 of last year showed that small pilchard tags ap proximately 13 mm. x 3 mm. x 1/2 mm. produced less mortality in tagging young herring than smaller but thicker tags inserted in the same way. In or^er to determine the efficiency of the electromagnets installed in reduction pla&i* in recovering small tags of various sizes, two1 lots of fish were tagged experimentally with small pilchard tags and flat #16 gauge wire tags 10 mm. and 7 mm. long and were thrown into the unloading chute at the Nootka plant. Returns were as follows:

Time of recovery in No. days after insertion: % Type of Tag inserted 1 2 3 4 Recovi

Small pilchard 10 6 2 1 mm 90 Flat wire - 10 mm. 10 4 2 1 X 80 Flat wire - 7 mm. 10 1 1 - mm 20 - 61 -

J.L. McHugh No. 44

The results show that the Nootka plant recovered a higher percentage of small pilchai'd tags than of the small wire tags. In view of the lower mortality associated with the flat type of tag, it was concluded that this type was most suitable for tagging young her ring. - 62 *

No. 45 J.L. Hart

The Pilchard Fishery in 1340 The pilchard fishery of 1940, could not be regarded as success ful. The total recorded estch. 6t. adiilt fish was slightly more than twenty-eight thousand tons} Thiis. was mbre than five times the catch of the previous year but did hot cokpare with those of 1938 or 1937 when comparable efforts were made in fishirigi In comparing these catches with those of still earlier years it should be borne in mind that the improvements in experience and equipment during the past few years pro bably increased this season's catch by thirty or forty per cent over what it would have been with the fishing Conditions of 1932. The catch per seine boat for the season was approximately 1175 tons. This is to be compared with the approximate figures. 210 tons in 1939 and 2150 tons in 1938. The fishing sSasdh Resembled that, of 1938 in the presence of two phases—a United States coast phase subceeded by a Canadian coast phase.. It differed from the previous season in the lack of a clear-cut distinction between the twb pb&ses»#jHshiiig was going on. simultaneously on the two fishing grounds for some three weeks, and in the comparative paucity of fish during the second phase 6ff the British Columbia coast. A minor difference lay in the concentration of fish well off the south eastern and central parts of the west coast of Vancouver Island rather than inshore at the northwest end of the. West coast. The future of the fishery is uncertain* For the immediate future the possibility of a good season exists but it does not appear to be good. The behaviour of the yoiing fish which were present on the fishing grounds last season ts impossible to predict. If they should join with the schools of adult fish which are being exploited fishermen will *e troubled by gilling to an important degree and plant operators will experience difficulties due to variety in the size of the fish to be processed. The long term prospects for the fishery appear to depend upon the 1939 year class of pilchards. If it is numerous enough to constitute a dominant year class after being exploited by whatever fisheries may exist for it as young fish during the next two or three years the pilchard supply may well be materially improved. Otherwise the indications are not promising.

J.L, Hart No. 46

Sampling the Pilchard Catch in 1940

Pilchard sampling was continued at Ucluelet during 1940. The sampling showed that the downward trend in length observable during the previous three years was continued. The average lengths for the years in question follow: - 63 -

J.L. Hart No. 46

Year Males Females mm. mm. 1937 248.0 252,0 1938 246.4 251.0 1939 241.0 245.4 1940 236,7 241.9

The decline in itself is not an indication of depletion or incipient failure of the fishery but the existence of a downward length trend in the fish supplying a fishery which is under suspicion of depletion and which is showing reduced catches for unit fishing effort is disquieting. The samples have been grouped according to fishing areas num bered from VIII to XI from south to north. Average lengths tended to increase from south to north with the more northerly areas indistin guishable .

Area Males Females

mm. mm. VIII 229.5 234,9 IX 232.5 238,2 X 238.4 243.6 XI 238.6 243.2

This almost exactly parallels the situation in 1938. The average vertebral count was 50.76. This is higher than any other season's average since sampling began. The significance of the change cannot be stated at the present time but the final elimination of the effects of the low-vertebral-number year class which entered the fishery in 1931 would appear to be indicated. Males were more numerous than female pilchards constituting 51.0 per cent of the fish sampled. During the previous season males were slightly predominant. The predom inance of males in these two years sets them apart from all previous seasons. In the earlier years males comprised about forty-five per cent of the pilchards sampled.

J.L. Hart No. 47 Application and Recovery of Pilchard Tags in 1940 Pilchard tagging in a curtailed form and pilchard tag recovery have been continued in an attempt to study the details of movements of pilchards. During the 1940 season 2600 tags were used, all of them off the Washington and Oregon coasts. Later in the season fishing moved northward to the waters off British Columbia but no tagging was done there as the supply of tags was exhausted. * *• u While tne fishery could not be termed successful the tonnage of fish captured was some five times as high as that of the previous year - 64 -

J.L, Hart No. 47 and opportunity for recovering tags was correspondingly improved. Ninety tags have been recovered since the preparation of the last re port. The accompanying table summarizes the results of pilchard tag recovery. The following points merit comment: (1) No returns have as

Tags considered Washington Oregon California British Tot( Columbia

Canadian tags 1936 0 1 0 1 1937 0 1 0 1 1938 0 3 0 3 1939 0 1 2 3 1940 8 0 30 38 Holdovers 0 2 0 2

Oregon tags — •» 3 3

California tags •» «• • - •» mm 39 39

Total 90

yet been received from Washington. There, the fishery was very poor and it is possible that no recoveries were made; on the other hand the fail ure to receive returns may be caused by some delay in the recovered tags being forwarded to the Biological Station. (2) After a year's absence from our recovery lists Oregon tags have reappeared. Biologically they are scarcely to be distinguished from Canadian tags used during the part of the fishing season when Canadian boats are working off the mouth of the Columbia River. (3) The large proportion of Canadian tags from the earlier tagging years as compared with 1939 tags recovered in California during the 1939-40 season. This may be contrasted with the recoveries by Canadian plants during 1940. (4) The large number of California tags recovered in Canadian plants. It is believed that the number is large because the California-tagged fish on Canadian fishing grounds have been augmented by fish tagged in 1937 and 1938 which in former years were too young to reach Canadian fishing grounds. This explan ation fits with the current hypothesis concerning the movements of pil chards. (5) The recoveries of a Canadian 1936 tag in California and a California 1936 tag in British Columbia are of interest in showing the length of time pilchards may remain in the fishery. The former had been out approximately three and a half years and the latter for nearly four and a half years. The ages of these fish when tagged are not known but they may be conservatively estimated at not less than four years. (6) The years of tagging for the California tags are as follows: 1936 1; 1937 —12; 1938 —13; 1939 — 13. It is proposed to supplement the work by making calculations to show whether during 1940 tags were more abundant in the early (United States coast) pilchards than in the later (Canadian coast) fish. This may throw light on the question as to whether all the fish on Canadian fishing grounds move north from California waters during the early summer, - 65 -

J.L. Hart No. 47

$n view of the uncertainty on this important point it is suggested that the tagging and tag recovery programmes be maintained in co-operation with United States federal and state authorities. The tagging of young pilchards is reported in another section.

J-L' ^ No. 48

Young Pilchards in British Columbia

During the summer of 1940 phenomenal numbers of young pil chards were present in British Columbia waters. As they were so numer ous and so widespread that it seemed possible that they might presage an economically important change in the abundance of adults, they were made the subject of an investigation. Attempts were made to obtain in formation by four methods: survey; canvassing fisheries officers, fisher men, etc. for information; sampling; and tagging. Survey. As the occurrence of small pilchards in British Columbia was so unprecedented it followed that no one in Canada could have sufficient background of experience to judge whether the abundance of young was sufficient to produce a dominant year class* Accordingly an attempt was made to have Mr. J.B. Phillips of the California State Fisheries Labor atory make a survey. (Mr. Phillips has made two surveys of the kind in California and Mexico waters and is undoubtedly the best qualified per son for the work.) Although preliminary arrangements were completed and tentative plans for transportation were made with the Canadian Depart ment of Fisheries and the United States Bureau of Fisheries, the arrange ments could not be completed owing to regulations governing the activi ties of employees of the State of California. Canvass for information. Through the intermediation of the Fisheries Department and directly, fisheries officers and those engaged in the fishing industry were questioned concerning the occurrence of young pilchards. The fish were first noticed and recognized in Barkley sound in the first week of May. They were sampled there by Mr. L. Quickenden on May 17. who found the different schools to contain considerable but widely varying proportions of anchovies. Our first knowledge of the presence of pilchards on the east coast of Vancouver island was our own observation at Departure bay on May 20 where also they were sampled. From reports obtained subsequently it would appear that by that date young pilchards had spread over most of the Georgia strait area although no reports at hand indicate their presence in. the Pender harbour area until considerably later. On May 28 the young fish were reported from Nootka sound and from all other fishery areas on the west coast south of Cape Cook during the same week. No young pilchards were reported to us from Quatinso sound until the week ending June 15. However, reports from were a week earlier. Later reports record the fish - 66 -

J.L. Hart No. 48 from Queen CharlQtte sound, Milbankt sound, Bishop bay, and Surf inlet (Sept. 18) on the mainland coast and at Pacofi and Selwyn inlet on the Queen Charlotte islands. During much of the summer young pilchards were reported as abundant off-shore off both the Canadian and United States coasts* On the west coast of Vancouver island pilchards were reported as becoming less abundant by the end of August. Negative reports are not at hand for all areas but it is believed (Nov. 25) that the schools of young fish are not being observed at the present time. In contrast to the west coast condition young pilchards are still being observed on the east c©a%t of Vancouver island. No data are at hand concerning present conditions in the more northerly areas. Considerable interest attaches to reports from three sources of pilchards being stranded and killed in fresfc water creeks and on tide flats. Sampling. Eight samples of young pilchards were examined from several localities^ The results set forth in the following tabulation give a general indication of the course of growth in the young fish during the year. It will be seen that while length has increased by approximately sixty per $enfc weight increased by two to three hundred per cent. Scales from the fish in these samples have been examined and it has been deduced from them that the fish were in their second year, i.e., belonged to the 1939 year class* The scales have been sent to Dr. L.A, -tfalford of the United States Fish and Wildlife Service who is making a specialized study of pilchard scales for examination and further study. He has already con firmed our age diagnoses. The samples after examination (in most cases) have been sent, to Dr. Frances N. Clark for racial investigations which will principally depend upon vertebral counts.

Date Source of sample Length Weight 1940 Central tendencjr average mm. grams

May . IT Kildonan harbour and Vernon 95 — bay, Barkley sound May 20 Departure bay 106 13 June 17 Green cove, Barkley sound 103 10 July $1 Barkley sound 120 17 July 28 Oregon coast 138 36 Sept* 10 Cowichan bay 146 34 Sept. 13 Spiller channel 133 34 Nov* 23 Nanaimo harbor 155 42 Tagging. One thousand young pilchards were caught with a bait seine in Cowichan bay on September 10. They were tagged as carefully as possible and released within a few hours. Returns from this part of the inves tigation are not expected within the next two years. m* 67 *»

J.L. Hart No. 49

Sampling Ling Cod. Fishing in oonnection with the general tagging work on the strait of Georgia has been conducted so as to catch as many fish for tagging as possible. For this reason the fish so caught provide a re presentative sample of the fish to be taken by the jigging method, Ool^ lection of. the available data shows that lengths between six hundred millimetres and six hundred fifty millimetres are most commonly repre sented. No fish shorter than 370 mm. were taken in our fishing and it is evident that the gear is selective against fish shorter than five hundred fifty millimetres. It is evident that sampling in this way gives a sample of the fish on the grounds as selected by the fishing method rather then a sample of the catch. The sizes of the fish in the catch will run considerably larger as the commercial fishermen are interested not in the number of fish caught but in the weight. A relationship between length and weight has been calculated as: Weight (ounces) w O.OOOO678 Length (millimetres) to the power 3*245. There has been considerable discussion of the advisability of applying minimum size regulations to the commercial ling cod fishery. Fishermen have suggested a minimum size limit of three pounds and other interests have been in favour of a higiher minimum size limit correspond ing to about five pounds. Using the weight-length relationship and the length frequency distribution the proportions of the unselected oatch which would be discarded can be determined. The proposed minimum size limit of three pounds corresponds to a length of 540 millimetres and would reject 17 per cent of the unselected catch. As the rejected fish would all be of small size the loss to the fishermen would be comparatively small and it is believed that such a regulation would be well received among all branches of the industry. The minimum size limit of five pounds corresponds to a length of 610 millimetres and would lead to the discard ing of 44 per cent of the unselected catch. As the discarded portion would have an average weight of more than three pounds per fish its loss would constitute a serious diminution of income for the fishermen. Some fishermen operating with inferior equipment might even bo forced out of business. Accordingly, although the highor minimum size limit is in many ways desirable, it is not recommended until such a time as conditions indicate that radical curtailment of catch is imperative or until bio logical work can demonstrate with assurance that such a regulation will work for the best interests of the ling cod fishery as a whole. The reoords for fish sampled in two areas of the strait of Georgia were kept separate. One of these areas was a place closed to commercial fishing (Saanich inlet), the othor was a place known to one fisherman only and fished by him alone so that it was exploited with re straint. A comparison of the lengths of the fish sampled in the three groups is of interest; -.68 -

J.L. Hart No. 49

Central tendency Percentage 1000 mm* Percentage 54t) mm. Lengths—mm. or more (c. 24 lb.) and less (c, 3 lb.)

General area 630 1.4 17 Closed area 900 28.0 3 "Private spot" 950 38.0 0

The "private spot" was a limited area including no shallow water which is usually regarded as the preferred habitat of the smaller fish. The maintenance of differences between freely exploited and restricted areas offers additional evidence for the lack of migration indicated by the tagging work. The nature of the difference serves to indicate the re action of the ling cod population to exploitation.

J.L. Hart No. 50

Tagging Ling Cod.

Since ling cod tagging was begun more than twelve hundred fish have been tagged. Of these some three hundred thirty were tagged by a modification of the standard method which has since been proved to be less effective. One hundred of the tags were used to the north of Vancouver island. The rest of the tagging was done around the strait of Georgia and the southern end of Vancouver island. The whole of the southern region has been divided into forty convenient arbitrary subdivisions. The proportion of tags recovered in different areas varies from none to twenty per cent (One tag only was used in one area and it was recovered making one hundred per cent). The highest proportions of returns have for the most part been from areas where the fishery is actively pursued. Tags put Out in an area dosed to commercial ling cod fishing have been recovered to a small extent only. Tag recoveries have continued to show that the majority of ling cod are recovered at the place where they were tagged or within a mile or two of it. There are, however, notable exceptions as was pointed out in the report for last year. One of the exceptions this year was 1kat of a fish tagged in October at William Head and recaptured four months later inside Cape Beale at the entrance to Barkley sound. The tagging at the south end of Vancouver island was carried out to investigate sug gestions that the gulf of Georgia ling cod supply was replenished from the and the ocean. So far no indications of a movement in the inward direction have been observed. Of the 96 tags which have been recovered and returned, 38 for which full data are available, were out for five months or more and are considered as being potential sources of information on the growth rate. - 69 -

J.L. Hart No. 50

Calculations based upon the weights and lengths at the times of togging and recovery and the interval of freedom after tagging show that ling cod grow at the average rates of 6 mm. and 3.0 ounces a month, or 2.8 inches and 2.2 pounds per year. Individual fish varied greatly, some of them actually losing weight, but there was no strongly marked relationship between rate of growth in either length or weight with size at the time of tagging. The smallest fish tagged and recovered did show the fastest increase in length (71 mm. in six months) but too great reliance cannot be placed on such a single record. In these calculations all months of the year have been regarded as of equal value in promoting growth and the sexes have been considered together. Until more data are available more refined calculations do not seem feasible. In continuing the ling cod tagging programme it is hoped that more fish may be tagged. As time is available it is proposed to make further oontacts with the fishermen in an attempt to improve the number of recoveries made and the accuracy of the information provided with them. . 1

-70-.

J.I). Hart and J.L. McHugh No. 51

Life-histories of British Columbia Smelts.

Information on the life-history of the four speoies of Osmeri- dae found in British Columbia waters is being collected whenever possible. Samples have been obtained.through the co-operation of officers of the Department of Fisheries, and also by the investigators themselves. The eulaohon and surf smelt support fisheries which are of considerable sea sonal importance. They and the capelin are taken in some numbers by the local inhabitants during the spawning run.

Eulachon - Thaleichthys pacificus This species ascends most of the larger rivers along the coast on its spawning migration during April and May. The majority of the fish are apparently in their second year, although a small proportion of threes is present. Males are apparently considerably more numerous than females, have a slightly lower vertebral count, and are readily distinguished by their secondary sexual characters. Comparison of average vertebral counts shows that free intermingling between runs to various rivers does not take place. Spawning in the is described elsewhere in these reports (McHugh). Little is known of the marine phase of the life—history. Specimens taken in Barkley sound and off the islands in the gulf of Georgia were found to be feeding entirely on euphausids.

Capelin - Mallotus catervarius The capelin is also seen in great numbers only on its spawning migration, which occurs during September and October. All fish appear to be in their second year. The secondary sexual differences are very marked at spawning time and significant differences in vertebral count can be demonstrated between males and females. As in the eulachon, males are apparently more numerous than femaleso Comparison of vertebral counts of runs from Nanoose to Departure bay gives no indication of a tendency to form local populations. Spawning takes place in the surf at the water's edge, and occurs generally at night. The eggs become attached to particles of coarse sand and take approximately two weeks to hatch. The larvae disappear soon after hatching. The food of two individuals taken from salmon stomachs was com posed entirely of euphausids.

Surf smelt - Hypomesus pretiosMs In British Columbia waters this species generally appears on its spawning migration during the period from June to September. Several age. groups may be found in the spawning populations, some fish maturing at the end of their first year, and others being in their second or third year of life. Males are more numerous than females and are smaller in average length. The external differences between the sexes are not so definite as in the other species, and no significant difference in vert ebral count can be demonstrated. The runs to various beaches tend to form local populations. - 71 -

J.L. Hart and J.L. McHugh jy0# $i

The spawning process is similar to that described for the capelin, and the eggs hatch in approximately the same space of time. The young fish have been observed in small schools in shallow bays among the ©ilf islands during the early summer months. Older fish with im mature gonads taken in Esquimalt harbour were feeding on euphausids, shrimp larvae, and tube-dwelling amphipods and worms.

Long-finned smelt - Spirinchus dilatus This species, which is the least common of the four, spawns a short distance up freshwater streams in the period from October to Decem ber. Specimens are sometimes found in shrimp nets, and a small sampie taken by this method in in March was apparently beginning its second year of life. In this sample females predominated and averaged 110 mm. in standard length. Breeding males can be distinguished from females by their secondary sexual characters. Four individuals in their first year, taken in May in the Fraser river, had been feeding entirely on Neomysis mercedis. The larger fish in the sample from Burrard inlet had been feeding on euphausids, while the smaller specimens had eaten copepods and occasional cumacea.

J.L. McHugh No# 52

Ova Diameter Measurements in Smelts.

Eggs of various sizes may be found in a maturing or ripe smelt ovary, ranging from the small transparent immature stages which are in visible to the naked eye to the yellowish, translucent, ripe ova which break from the follicles and lie free in the lumen of the ovary. By con structing frequency polygons based on the diameWs of all eggs found in a small section of the gDnad it is possible to deduce to seme extent the spawning behaviour of the fish. In the eulachon, measurements made on ovaries taken at the beginning of the run, and approaching the ripe oondition, show two modes on the diameter frequency polygon, one at 0.135 mm. representing the im-- mature eggs, and one in the vicinity of 0.95 mm. which represents the ripening eggs. Measurements from a spent overy show a similar mode at 0.135 mm., one near 1.04 mm. representing the remaining unspawned ripe eggs, and a more or less regular series of intermediate stages, rising to a mode which also lies at 0.135 mm. These latter are unspawned ova in various stages of degeneration and reabsorption. The curve for spawning capelin contains at least three modes, one of immature eggs average 0.135 mm. in diameter, one of mature eggs near 1.0 mm., and at least one other somewhere between these two, which represents a second group ripening out from the immature class. In the surf smelt it has been shown by other workers that suc cessive groups of eggs ripen out from the main immature mode as the season -72 -

J.L. McHugh No. 52 progresses. Examination of ovaries from samples taken in British Colum bia waters shows the presence of modes at various intervals between the immature and the ripe ova. It is concluded that the two speoies of smelt spawning in salt water ripen successive batches of ova from the immature group, and there fore spawn more than once in a season, while the eulachon apparently spawns only onoe.

J.L. McHugh No. 53

Eraser River Eulachon.

At the request of the Dominion Department of Fisheries the investigation of the life-history of the eulachon, commenced in 1939» was continued during the current year. Samples and information were again obtained through the courtesy of the Fisheries officers, and a visit was made to the Eraser river to study spawning. According to reports reoeived, the run of eulachons in all dis tricts was considerably heavier thr.n average during the 1940 season. Re ports that the fish were larger than in previous years were borne but by comparing samples taken in the Eraser river in 1939 and 1940. Samples were examined from five districts: the Nass river, Rivers inlet, Kingcomo inlet, Knight inlet, and the Eraser river* In average length the fish from the Nass compared olosely with those from the Eraser, but the fish from the central area were considerably larger. Al though the determination of ago by the scnle method is difficult for this species, it nppeared that the bulk of tho run to each locality was com posed of fish in their second year, and thereforo the difference in size was not duo to differences in ago composition. Average vertebral counts of samples from e^oh locality wero As follows: Vertebral Number Number of Fish Locality Males Females Males Females

Eraser river 67.782 68.107 142 56 Knight inlet 68.236 68.382 72 34 68.293 68.526 99 38

Rivers inlet 68.202 ~ 89 8 Nass river 68.563 68.492 119 65

In all but one locality females were found to have a slightly higher count than males, Comparing vertebral counts of males statistically, it was found that significant differences could be demonstrated between the Nass, the Eraser, and the central area. It was therefore concluded that the runs to these areas tended to form local populations between which free inter mingling did not take place. A gradation in vertebral count with lati tude was observed, fish from northern waters having the highest counts. - 73 -

J.L. McHugh No. 53

By means of an extended series of dredge hauls taken in the Eraser river, it was discovered that spawning takes place in the river bottom in the vicinity of Mission. The eggs become attached to the sand particles by means of a short peduncle formed from the outer membrane. Spawning apparently occurs only where the sand particles are of a certain specific size, and this limits the areas of river bottom in which eggs may be deposited. Only newly-hatched larvae were found in the river it self, and it was concluded that the young fish are carried to sea shortly after hatching. For the proper regulation of the fishery it is necessary to secure reliable information on the relative strength of each year's run. It was recommended that this information could best be obtained by the collection of adequate catch statistics. - 74 -

t,D.B.-^ Quayle~. i No. 54

Natural Oyster Propagation in Ladysmith Harbour

Japanese oysters -^ Japanese oyster propagation was again a failure in 1940. High temperatures during the spring and early summer caused the oysters to ripen very early. As a result spawning took place com paratively early in the summer. Some spawning occurred on June 26 and small lots of "wild oysters spawned at intervals from then until July 16, when an exception ally large spawning took place on one of the commercial beds with water temperatures around 23°C, On July 18 numerous straight-hinged larvae. appeared in the plankton from this latter spawning. The plankton of July 19 showed larvae of three different age groups, the older ones ap parently originating from the small spawnings of "wild" oysters prior to July 16. By July 26 there were still considerable numbers of larvae in the plankton, consisting of four size groups, the largest of which were two-thirds grown. But by August 5, all larvae had completely disappeared. On August 6 another large spawning occurred and the young larvae were found in the plankton on August 9. On August 16, examination of all the beds in the harbour showed that 90$ or more of the oysters were spawn ed out. On August 21, the larvae in the plankton were still fairly nu merous but by August 26 only one or two of them, two-thirds grown, could be found. No other Japanese oyster larvae could be found in the plankton during the remainder of the summer. Weather and water temperature conditions during this summer were as good or better than the average and all the oysters in the har bour spawned out satisfactorily, without aid by stimulation. As ifar as could be-ascertained, larvae were produced from all spawnings and con siderable numbers grew to be two-thirds the normal spatting size or near ly two weeks old before disappearing from the plankton. Several larvae were observed that were almost ready to spat, but none was found on the spat collectors. No measured physical or chemical conditions could be correlated with the occurrence of larval mortality.

Native oysters

The native oyster industry in British Columbia still remains in the doldrums with little commercial activity anywhere. During the summer of 1940, the native oyster propagation in Ladysmith Harbour was also a failure. Larvae appeared in the plankton but no extensive spat ting occurred. No explanation can be offered. Too little is kno?/n about the fundamental biology of these animals, especially the larvae, to form any conclusions regarding failure to spat in 1940. - 75 -

D.B. Quayle No. 55

The Pond Propagation of Japanese Oysters

The Japanese oyster industry in British Columbia is heavily dependent upon the shipment of oyster seed from Japan. If that supply should fail the industry would suffer very seriously. Therefore it seemed desirable to attempt another trial of propagation in Ladysmith harbour. Several spawnings have taken place in this area since 1932 but except for the unusual "sets" in 1932 and 1936, spatting has been neg ligible, even though temperature conditions have been equally as favour able as in the two successful years. In experiments conducted at Conway, England, successful sets of the European oyster have been obtained in large cement mussel puri fication tanks by enriching the organic content of the water and in creasing the nannoplankton stock. Therefore it was decided to test a somewhat similar method at Ladysmith but, in this case, using a small controllable lagoon at low tide approximately 3.6 acres in size, with water at an average depth of 4 feet. During the summer of 1940 a dam 90 feet long, 27 feet wide and 10 feet high, largely of earth but with a cribbed outlet, heavily rocked, was constructed. The opening had to be strongly made because an aver age of a million gallons of water per hour passes out of the lagoon dur ing the spring high tide period. Most of the work was done by the author and an assistant. The plan for 1940 was to dam up the entrance of the lagoon, introduce a spawning stock of oysters, stimulate spawning and the pro duction of larvae in the lagoon, analyze the nannoplankton to determine the presence or absence of what might be suitable food organisms and, finally, provide organic enrichment to the water, if necessary. Unfor tunately by the time the dam was completed, August 11, all the planted stock in the lagoon, as well as mo£=t of the oysters in the bay had spawned. The experiment, therefore, could not be made. At the end of two weeks the sluice box in the dam was opened and the tide allowed to flow in and out of the lagoon in the normal fashion. From the experiences during 1940, however, it was found that - 1. The average water temperature in the pond after damming was 3.6°C. higher than the average temperature recorded at the thermograph station in the harbor. 2. The salinities remained quite constant near a level of 25.5 °/oo which is considered the optimum for the development of Japanese oyster larvae• 3. The oxygen concentration was in the main, quite high although it took a sharp drop for a period of two days after which it rose to its former level. Beyond the possibility of it being caused by the oxygen consumption of decaying Enteromorpha, the cause of the drop is not known. 4. About two tons of oysters were induced to spawn with little difficulty. No larvae, however, appeared in the plankton. -* 76 -

D.B. Quayle No. 55 5, Native oyster larvae and numerous polychaete larvae were found in the plankton, indicating that conditions were not too abnormal, Mr. H. Carmichael the owner of the property on which the dam was built, kindly gave permission to use the pond and to construct the dam. Financial assistance was received from the British Columbia Packers, Ltd., and the Crescent Oyster Company,

D.B. Quayle No. 56

Condition Factor in Oysters

In conjunction with the Washington State Pollution Commission which was studying the condition of oysters in that State, tests were made of samples of oysters from Ladysmith harbour and . The technique is as follows: the oysters are sponged with a towel to remove extraneous water, then placed in a glass "graduate" and the amount of displacement of water noted. The oysters are then opened, the "meats" allowed to drain for ten minutes and then the volume of the "meats" and of the empty shells are separately measured, by displacement. The condition factor is found by dividing the volume of the meats by the inside volume of the oyster, i.e., the volume of the whole oyster minus the volume of the empty shell. Four lots of samples of 100 oysters each were taken from Ladysmith and five from Boundary bay in March, 1940. The condition fac tors were found to be as follows:

"Condition" Factor

Sample No. Ladysmith Boundary bay

1 .4729 .4500 2 .4545 .5172 3 .4484 .4861 4 .4563 .4502 5 .4866

Average .4580 .4799 In January, 1940, the condition of Puget sound oysters varied between 0.490 and 0.416. As was expected, the Boundary bay oysters showed a slightly higher condition factor varying from 0.450 to 0.517 with the average at 0,48. Ladysmith indicated a narrower range of 0.448 to 0,473. with an average of 0.458. This method of establishing a condition factor for oysters should, with certain refinements, prove fairly speedy and very useful. - 77 -

D.B. Quayle No. 57

Statistics of the Commercial Clam Catch

In order to obtain information as to the annual catch of each species of clam, the effort expended, the annual fluctuations in abun dance, etc, a system of catch statistics was put into practice in the 1938-39 season. The scheme was somewhat modified in 1939-40 to make the details of catches more readily usable. From the data submitted on the official clam report forms, it is possible to compute the following: 1. Digging intensity, i.e., the number of man-tides, or the effort expended on any one beach or area, 2. Quantity of clams produced from any one beach or area, 3. Average catch per man-tide for any one beach or area. The catch per unit of effort assumes that the fraction of the total population taken by the fishery is proportional to the effort ex pended by the diggers. Thus, in any one year, the average catch per unit of effort is a constant fraction of the clam population existing on the beaches in that year and as such may be used as a measure of the abundance of clams in that year. The following table summarizes the records for the "butter" clam, the species normally used for canning, for the 1939-40 season. These records represent about 60 per cent of the total catch as shown by the records of the Department of Fisheries. The areas listed are designated by the location of the buyer, the first three all being in the Alert bay area. The total catch is obtained by summing all the de liveries from all the beaches in the area. The total man-tides is the product of number of diggers times number of digging tides, for all beaches in each area.

Area Total Catch Total Average Catch (lbs.) Man-tides per Man-tide (lbs.)

Old Village 584,000 2780 210.1 Alert Bay 246,270 1426 172.6 Chatham Channel 556,260 2704 206.2 Chemainus 202,050 1684 120.0 Sidney 181,080 1312 138.0 - 78 -

D.B. Quayle No. 58

Clam Productivity as Tested by Latin Square Plot Experiments

To test the productivity of various clam beaches, as well as the most efficient method of exploitation, Fisher's method of plot ex perimentation utilizing the Latin square, was begun in 1938. The experiment involves five treatments as follows: Plots to be dug (1) twice a year, (2) once a year, (3) every other year, (4) every third year, (5) only at the end of the experiment. Plots are located at seven strategic and typical beaches on the east coast of Vancouver island from Knapp island (near Victoria) to Echo bay (Alert bay). The experiment is now entering its third season when treatment number three will be applied for the second time. Shortly sufficient data should be available to permit preliminary analyses of results.

D.B. Quayle * No. 59 Sampling the Commercial Clam Catch

To determine the length and age composition of commercial clam catches and the extent of annual variations in these characters, sampling of the commercial clam catches during the 1939-1940 season was undertaken. Three areas were investigated, Sidney, Chemainus and Alert bay. Thirty-seven samples of 100 clams each were taken during the season. To date the data have been completed only to the frequency dis tribution stage. Consequently no conclusions can yet be derived. The clam industry cannot be regulated by the individual beaches but rather only by combining the number of beaches into a definite area. Therefore sampling at a cannery when the origin of the clams is known only in terms of areas is considered adequate enough to give a fairly true picture of the size and age composition of the commercial catch.

D.B. Quayle No„ 60 Growtl. Experiments on Commercial Clams

Growth experiments on the two most important commercial species, (Saxidomus giganteus - the butter clam, and Paphia staminea - the li ttle neck), instituted in 1939, has been continued and enlarged. The purpose of the work is to determine the seasonal and annual growth in various localities and at different tidal elevations and to test the validity of the "annual ring" method of age and growth determinations. Marked clams, bearing engraved numbers, are utilized for the ex periments and are retained under differing natural conditions in suitably situated wooden boxes. The field work on this problem will be completed in the spring of 1941, From seasonal examinations, much data have already been collec ted but not analyzed. Upon conclusion of the work, a report will be pre pared. - 79 -

D.B. Quayle No. 61

Experiments on the Burrowing of Clams

Tests were made to determine whether, after a clam bed has been dug commercially, those individuals below legal size and left on the beach would burrow back into the sand and gravel of the beach. Only the hard-shelled clams were involved. Four wooden boxes, 4 feet square and 3 inches high, open at the bottom and covered with wire netting at the top were constructed. On the clam beach at Departure bay a plot 8 feet square was laid out and divided into four equal parts. Two of these were dug over and all clams encoun tered removed. These dug-over plots were termed "soft" ground and sim» ulated conditions in commercial clam diggings, whereas the unbroken plots were designated "hard" ground. One hundred recently-dug "butter" clams were placed on one of the "soft11 ground plots and another hundred on a "hard" ground section. One hundred recently-dug "little neck" clama> were placed in each of the two remaining plots. The whole area was covered with the wire-toppe* boxes to prevent loss of clams by wave action or removal by ducks, etc. The plots were examined at the end of twenty-four hours to discover how many clams had burrowed back into the beach. The experiment was repeated four times at roughly three-month intervals to note any possible influence of seasonal air and water tem perature conditions. In summer the tides are low during the day, hence the clams are exposed to high air temperatures during low-tide periods as well as to high water temperatures. In winter the low tides occur at night, consequently the clams would be subjected to low air and water temperatures. The results for the four experiments are tabulated below and show the number of clams that had burrowed back into the beach by the end of a 24 hour period.

No. Date Little n eck clam Butter clam

Hard ground Soft ground Hard ground Soft ground

Jan. 21, 1939 100 98 97 93 April 13, 1940 98 88 92 90 3 Aug. 3, 1940 96 91 55 63 4 Nov. 19, 1940 95 96 74 85 It is apparent that both species readily burrow back into the beach and, in the main, do so relatively soon after being turned out. Little distinction between "hard" and»soft<- ground is indicated, except perhaps for the "butter" clams in the last two tests. These-two samples also suggest that summer temperature conditions may have some effect. This latter possibility deserves further study. - 80 -

D.B, Quayle No. 62

Reproduction in Clams The histological work and general observations on the repro duction of the commercial species of clams has been continued. The pur pose of the investigation is to determine the size and age at maturity* the spawning season, the time and method of gamete formation and the most suitable period for exploitation in regard to edibility. Samples of clams are taken from the beds at various intervals throughout the year. Slide preparations or "smears" of a portion of the gonads are made. The study is proceeding satisfactorily for both "little neck", Paphia staminea and "butter" clams, Saxidomus giganteus.

D.B. Quayle ^ qq

A Quantitative Investigation of Bivalve Larvae

In 1939 a quantitative plankton investigation was undertaken in Ladysmith harbour and Departure bay to obtain definite information on the spawning periods of the various bivalves and to determine the growth rate of the larvae and the length of larval life. During 1940 the Ladysmith collections were continued. Six permanent stations were occupied bi-weekly and ten minute surface tows (0 to 3 feet) were taken. A plankton net of No. 25 bolting silk, with a ten-inch diameter and a length of thirty inches was used. The Depar ture bay collections were stopped, however, since the 1939 results were too meagre to warrant further sampling. Most of the plankton samples have been counted but no analyses of the data have yet been made. - 81 -

J,P. Tully No. 64

The Program in Hydrography

As was discussed in summary reports for 1939, the general program has been concerned with the study of circulation systems in Georgia strait, Juan de E\jtea strait, approaches to Juan de luca strait, and . It seems unnecessary to reiterate the details of the investigation here. It was planned to assemble this material during the past year, to prepare several papers on the theoretical problems involved and to discuss the findings of those phases of the investigation that have been completed. However, the attention of the department has been so interrupted with other problems, that this research has not been able to make the progress anticipated. A theoretical study of the relation of the rate of mixing to velocity of flow in stratified water has been made and an equation of motion has been derived which may be applied in the highly stratified enclosed and coastal waters. The equation provides a means of calculat ing currents in these waters where considerations based on the stream theorems are likely to be erroneous, but is not applicable in the deep sea where the latter may be used to advantage. The hydrodynamic calculations, and stream diagrams have been prepared for all series of data, but there remains considerable work to be done in the analyses of the dynamic systems, tidal, seasonal, and weather, that occur in the area. It is expected that some of the material will be available for publication next year. The primary activities of 1940 have been concerned with the continuation of the program of daily soa water observation, prediction of sulphite liquor pollution in Alborni inlet, investigation of saloon T mortality in Cowichan bay, and the routine sorvice work of chemical, physical, and engineering natures.

JVP. Tully No. 65

Daily Observation of Sea-Water

Following the plan of previous years, the system of observing the temperature and salinity of the sea water and comparing these to the meteorological conditions has been continued. It was planned to prepare some of the back series for publication in the form of a bulletin, with a view to establishing an annual publication of these data for the use of the fishing industry and such investigators as might require the information. It has not been possible to complete this project, but it is hoped that these data may be made available during the coming year. The systems of investigation may be briefly outlined as follows: Cases of bottles are sent to each observing station, where the observer takes a surface sample of sea water each day and notes the temperature. Twhen a case is filled (about three months) it - 82 -

J«P. $ully No. 65

is returned to this station where the salinities of the sea water samples are determined (by titration), the bottles washed, etc., and returned to the sampler. Co-incidentally the meteorological records of precipitation, air temperature, cloudiness, and wind direction and force are obtained from the Meteorological Service. These data, representing two sea- water and five atmospheric observations from fifteen stations on this coast, are smoothed by seven day running averages to eliminate the lunar cycles, and are plotted as a function of time. The resulting picture indicates the progress of the various sys tems in a readily comparable form. More labour is involved than is usually credited but it. is evident that in order to present a compre hensible comparison of so many factors it is necessary to simplify their daily fluctuations and yet preserve their continuity and trend. Henoe it is necessary to prepare the tables of running averages.

J.P. Tully No. 66

Alberni Inlet Investigation

As previously reported it Was desired to prognosticate the state of pollution that would result in Alberni inlet if the total ef fluent from a proposed 135 ton sulphite pulp mill, to be located at Port Alberni, were discharged into the inlet. A tidal current survey was made in which the water movements were determined with a direction finding current meter at a number of intervals between the surface and 100 meter depth. The probable concentration, and rate of transport of the effluent was calculated from these data. These etudies did not give any information as to the rate of exchange of water between the head and body of the inlet, consequently it was impossible to deteimine the extent to which the effluent would Accumulate in the head. A hydraulic model of this latter area was there fore built, in which tidal action, land drainage, wind and density gra dient, were simulated. The nature and time of accumulation of the mill effluent were studied in the model, a description of which has been pre pared and will be presented for publication. Numerous consultations of the literature and with persons acquainted with the subject led to the tentative adoption of an arbitrary standard of harmful pollution, and the conclusions of these investigations were presented in "Report on Proposed Sulphite Pollution of Alberni Inlet". It was indicated that the mill could operate and waste its effluent in the inlet provided that: (1) The effluent was greatly diluted and impounded for a time before entering the head of the inlet, or (2) rapid and com plete mixing occurred with all of the water in the head of the inlet; or (3) the effluent was piped seaward from the head of the inlet and re leased near one shore. It was further recommended that an investigation be made to determine the state of the inlet before and after the mill was - 83 -

J.P. Tully No. 66

constructed so that the degree of pollution could be determined in terms of the fundamental characteristics of the area. This would form the basis for prediction of effects that might be expected from other indus trial developments. This investigation is of particular interest for all avail able references indicate that the only studies of pollution that have been made were instituted after the damage had been done and there is no knowledge of the original state of the systems. Further comments on this investigation are being submitted to the Board by special report.

J,P. Tully No* 67

Salmon Catches and Physical Oceanography

During the past four years a number of trolling fishermen have been submitting a confidential record of their fishing effort, including time, location, water temperatures, weather record, and catch of spring and coho salmon. There has been a current belief for some time that a relation might be found between the physical conditions and the catch. There is the possibility of four major factors affecting the catch, namely, - (1) presence of fish, (2) locality of fishing, (3) in dividual characteristics of the fisherman and his gear and (4) weather and oceanographic conditions. By dealing with the average catches throughout the fishing season of a number of fishermen in a small area, the first three factors were reduced to approximately constant values. The catch per h*>ur o-f the group was then determined under the following conditions:

(Light

Strong

Rising

Maximum Water temperature Falling

Minimum

This provides six conditions of wind and four of water temperature. - 84 -

J.P. Tully No. 67

ssoatchfswJ4eXamPl6S- r<'=»»»°f the relationsunsbetween winaa/siwnaittcm.srand

Catch per hour

Lightt< v* f^S.E. 1.111,Q5 1.12Q>-96 6.662-62 3-842.96 0.28°'45 °'861.65 0.40°*™ 0.710.70 Var. 2.68 1.99 1.92 2.19 0.21 1.09 0.72 0.&7 West 2.56 1.47 2,32 2.12 0.37 0.83 0.24 0.48 Strong S.E, 0,15* 0.30 1.00* 0.45 0.48*1.14 0.60* 0.74* Var. 6.96* 3.50 0* 3.49 0.07* 1.04 0* 0,37* Data not significant due to small number of occurrences

More extensive analyses have of course been carried out than are represented in this paradigm, and the data have also been analysed for a relation between fish catch and water temperatures, tidal range, lunar phase etc., and a combination of two factors such as wind and tem perature but there is no apparent relation between any of these and the catch m any of the four areas, studied. On this basis it is concluded that the catch depends on one or more of the first three factors. That is, if the fish are present in the locality they may be caught by a skilful fisherman with proper gear at any time, provided that weather permits of the fishing operations, but weather conditions in themselves apparently do not affect the avail ability of the fish. It is probable that each troller's record may be of personal value as a guide to location of the best fishery and the most suitable type of gear, but it does not appear that any general relation can be found between the physical conditions and the fishery. Consequently there is no object in continuing this investigation. A detailed report is in preparation which will be circulated among the fishermen for fur ther discussion. «- 85 -

R.E. Foerster and John P, Tully N0# £g

Mortality of Salmon in Cowichan Bay.

From September 14 to 16 a heavy mortality, estimated at about 500 fish, was reported to have occurred amongst adult spring (Chinook) salmon. The mortality continued through.the ensuing fortnight at a diminishing rate. Following a thunder storm on September 12th, there was a period of bright calm weather, during which an outburst of diatoms (Coscino-* discus) occurred. Zooplankton were conspicuously absent and while no young fish were found dead, it appeared that these, along with the coarse fish and the herring, had left the bay altogether. There were no ap parent abnormalities in the dead fish, which were found to be edible. The water conditions, except for the pH were found to be quite normal. The high values of this property indicated a complete lack of carbon dioxide and bicarbonates. When the fish were observed returning to the bay, two weeks after the first mortality, it was noted that the pH had fallen' to normal values. By process of elimination of causative factors, as discussed in the current issue of Progress Reports (no. 46) it was concluded that the mortality was related to the high alkalinity. This condition may have so reduced the carbon dioxide tension of the water that the transfer of oxygen through the gills of the fish, to the blood, and from there to the tissues, may have been so disturbed as to cause death. There is very little information on which to base this conclusion and it is suggested as a possibility based on deduction and elimination, rather than n conclu sion based on established reactions. - 86 -

D.C.B. Duff No# £9

Furunouloais Investigation in 1939-40,

Introduction. In view of the results to be reported below, opportunity is here taken to review the considerations which led to initiating this investigation. (a) There is ample proof in the literature (1, 2, 3) that anti- genically unaltered protein may at times penetrate the normal mucosa of the alimentary tract in mammals, thence reaching the blood and the tis sues concerned in antibody production. The laboratory evidence of such protein passage has most frequently been the demonstration of hyper sensitivity in the affected animal, Ratner and Gruehl, (2) in review ing their extensive experiments, conclude that passage of unaltered protein occurred in guinea-pigs eo frequently that it could be regarded as a normal phenomenon, (b) The amount of unaltered protein absorbed as a result of feed ing is small, representing a very minute fraction of the total protein ingested. The bulk of the protein is, of course, enzymically hydrolyzed to amino-aeids, which by themselves are not serologically active, and which retain none of the immunological specificity attributable to the parent substance. However in the case of bacterial proteins, the exposure of animal tissues to a small primary antigenic stimulus usually alters the reaction of the animal to subsequent contact with the same antigen (e,g, an actual infection). This alteration is frequently in the dir ection of a much stronger immune reaction. The primary dose used in the production of diphtheria antitoxin in horses by means of toxoid inoc ulations is a standard example of this possibility. (c) Again, numerous reports, including 4,5,6, aotnally indicate that under oertain conditions and with certain antigens, the feeding of antigens to humans and animals, or the exposure of the human or animal nasal mucosa to antigenic preparations, produces a definite effect in increasing resistance to infection, and very frequently a significant rise in circulating antibody. (d) There is no evidence to suggest that the alimentary mucosa of fishes is less permeable than that of the higier vertebrates. The op portunity moreover exists in hatchery work of easily administering a much larger number of doses than would be practicable for human immunization work.

With these facts in mind it was deemed reasonable to attempt an alteration of reactivity to B. salmonlcida (furunculosis) infection in the cutthroat trout, with some expectation that the altered reactivity would tend in the direotion of increased immunity. (The fact that the cutthroat trout will react to injected doses of B, salmonioida vaccine, with the production of agglutinins and increased resistance, has already been shown.) It was indicated in last year's Summary Report that stocks of fish had been treated by oral administration of vaccine, but that the testing of these stocks had been prevented by the destruction of the Stan ley Park laboratory by fire in February, 1939, I am now happy to report - 87 -

D.C.B. Duff No. 69 that our new quarters are giving every satisfaction, and that vaccinated fish have been tested, and continue to be tested, without interruption.

Methods.

Vaccine. The vaccine used consists of a standardized, heavy saline sus pension of B, salmonicida. The growth is washed from agar surfaces with sterile physiological saline, and killed by saturation with U, S. P, chloroform (7), for 48 to 72 hours at 20 deg, 0, At the end of this time the chloroform is evaporated off, and suitable tests for sterility are conducted. _The preparation is administered by mixing it intimately with the regular hatchery food (containing a large proportion of beef liver). It may be remarked that much of the antigen is probably lost due to the relatively crude method employed. Better methods have been con ceived during the course of the work, and should be tried in the future. At the same time, the method used would appear to have produced definite results.

PROTECTION TESTS; Two (at present 4) thermostatically controlled wooden tanks were used, each having a capacity of 20 or more trout. Treated (orally vaccinated) fish, and untreated control trout were removed from the large, stock tanks and placed in the test tanks. In some experiments treated fish were placed in one tank, controls in another, under identi cal physical conditions. In other experiments treated and control fish were combined in both test tanks, and distinguished by marking the treated fish. Each tank was then infected with a standard amount of virulent B* salmonicida suspension. "~ A total of 9 teste have been carried out to date, and further tests are at present under way. Three of the 9 tests are not reported because of the occurrence in them of several deaths not specifically due to B, salmonicida infection, and in which it was therefore, not possible to draw conclusions as to the ratio of specific deaths in the treated and control groups. In the remaining 6 experiments» Table I, all deaths reported are "specific", that is, B, salmonicida was recovered from kidney'and/or heart blood. • 88 ~

D.C.B. Duff No. 69 TABLE I.

Protection Tests Deaths/Total Fieh f. Loss Test Control Test Control #2. 16 - 4 - 40 to 17 - 5 - 40 5/5 5/5 100 100 40 vaccine feedings prior to test.

5. 17 - 7 - 40 to 29 - 7 - 40 l/6 9/9 15 100 64 feedings

#6. 29 - 7 - 40 to 12 - 8 - 40 0/4 4/4 0 100 68 feedings. #7. 13 - 8 - 40 to 26 - 8 - 40 6/20 H/20 30 55 68 feedings.

#8. 27 ~ 8 - 40 to 19 - 9 - 40 6/20 17/20 30 85 70 feedings• mmMmmmmm*mmsaammmmmm—mtmmw—»-•—•—•-•ana«i^•_••> #9- 20 - 9 ~ 40 to 9 ~ 10 - 40 4/20 14/20 20 70

Totals 22/75 60/78 29.3 76.9 Totals omitting Expt. I, 17/70 55/73 24.3 75.3 Note: vaccine feedings reduced in number after Experiment $ 2.

The results recorded in Table I appear to demonstrate: (a) On the basis of a single test (No. 2), a probability was indicated that 40 doses of the vaccine used are inadequate to give protec tion. Unfortunately 2 non-significant experiments, involving 16 fish each, Intervened before sigiificant results were again obtained in No. 5. (b) 64 or more oral doses of vaccine confer a well-defined re sistance against infection. The percentage figures in the last column of Table I are not of course mathematically justified, but serve here as a convenient means of numerical comparison. Protection thus varied from lOOf. (Test 6) down through lesser percentages in other tests. The total losse among 70 vaccinated fish in Tests 5 to 9 inclusive was 17 (24.3f.) ocmpared with a loss in the controls of 55 out of 73 fish (75.3$). Fur ther tests along the same lines are at present being carried out. - 89 -

D.C.B. Duff U) vVW No. 69

TESTS FOR THE PRESENCE OF CIRCULATING ANTIB The agglutination tost was chps-eli fcaTthis purpose. It was found by preliminary experiment -ywnt^lmmune sera (rabbit) specific for B. salmonicida, that the most sensitive antigen (bacterial suspension) consisted of a 48-hour live growth washed from nutrient agar surfaces with 0.05?. saline, washed once by centrifugalization and resuspended in 0.05$ saline. The antigen if kept in the cold was suitable for use up to 48 hours after its preparation. Blood was obtained from the trout by means of cardiao exsanguination, the only method found to yield, in small fish, a sufficient amount of clean blood for test purposes. A small quantity of heparin was first injected into the circulation to prevent clotting. The standard agglutination technique was used, and final serum dilutions varied from l/lO to l/l60 (see Table HI).

TABLE H,

Agglutinin titres of S. clarkii sera

Date of Test From Vaccinated Fish From Control Fish Fish No, Serum Dil. Fish No. Serum Dil,

27-8-40 I 1 0 0 1 0 I 2 40 0 2 0 I 3 160 C 3 20

6-9-40 I 1 20 I 2 80 0 2 0 I 3 40 C 3 0 T 4 40 C 4 0 I 5 80 C 5 10 I 6 80 C 6 0

6-10-40 I 1 40 C 1 20 I 2 40 C 2 0 I 3 20 0 3 0 • I 4 0 C 4 0

23 - 10 - 40 • I 1 160 C 1 80 I 2 160 0 2 40 1 3 160 C 3 40

0 - no agglutination, 40, 80, etc. means definite agglutination at serum dilutions of l/40, l/80 etc.

J - 90 -

D.C.B. IXiff No, 69

TABLE III.

Analysis of Table II contents. Numbers of Trout Sera Possessing Specific Agglutinins to Dilution Shown

No Total Serum Dilutions 1/lC 1/20 l/40 l/80 l/l60 Agglut- Sera inins Tested,

Vaccinated Trout 16

Control Trout 9 . 15

The number of sera tested is still small and additional similar tests are at present being carried out. Within this brief series, how ever, there can be seen a distinct tendency (Table in) for more fish to show agglutinins, and for more sera to show higher agglutinin titres. Tha following figures are illuminating:

Ratio of Vaccinated fish whose sera showed any agglutinins 14/16 or 80.8f. Ratio of Non-Vaccinated fish whose sera showed any agglutinins .6/15 or 40.0?.

$ Ratio of Vaccinated fish whose sera showed agglut inins in dilutions of 1/40 or higher 12/16 or 75.Of, Ratio of Non-Vaceinated fish whose sera showed ag~ glutinins in dilutions of l/40 or higher 3/l5 or 20.0/.

Ratio of Vaccinated fish whose sera showed agglut inins in dilution of l/80 or higher 7/l6 or 43.8f. Ratio of Non-Vaccinated Fish whose sera showed ag glutinins in dilution of l/8o or higher 1/15 or 6.7^

Since individual fish have to be sacrificed to obtain blood for the agglutination test, it is impossible at present to obtain any direct correlation between resistance to infection and agglutinin titre. How ever with the completion of longer series of protection and agglutination tests, a statistical comparison should eventually be possible. - 91 -

D.C.B. Duff No. 69

Other evidence as to mechanism of protection A number of tests were made as to the ability of phagocytes plus serum of treated and untreated trout to ingest JB. salmonicida. Great difficulty was encountered in obtaining readable phagocytic smears, and the attempt was temporarily abandoned.

DISCUSSION; PROTECTION EXPERIMENTS. In spite of tjie relatively crude method of oral vaccine admin istration, definite evidence of protection has been obtained, through the use of a large number (64) of doses. Considering the averaged figures total losses were reduced, in the vaccinated groups, to about one third of tho figure for the control, non-vaccinated groups. There is little doubt that more efficient means of feeding tho vaccine can be developed, and that a more efficient vaccine preparation could be devoloped. Such developments might have the effect of reducing the total requisite num ber of feedings. Tho present number (64) may at first seem an unwieldy figure. But it is not an unpractical figure, since each dose, to the if hatchery man, means only adding the contents of a small bottle to the food at mixing time. Furthermore, daily doses may be given without harm to the fish. Thus if vaccine feeding were commenced as soon as the fry be gan taking food, the required number of doses would have been administered well before the commencement of warm weather, with its risk of Furunculo- sis. Preparation of a season's supply of vaccine would not be costly.

RECOMMENDATIONS; The relatively small number of experiments reported preclude, at the moment, any final discussion on tho true value of the protective measures under test. It would appear sufficient at this point to record the fact that a definite measure of protection is afforded, and to loavo a mathematically justified quantitative estimate of this protection until further data have been accumulated. Providing such an estimate is satis factory, recommendations would be made as follows: (A) Application of the system to hatchery practice. (B) Improvement of vaccine as to preparation and cost? improvement of feeding method.

Other Investigations. Further work has been carried out at the Univorsity of British Columbia laboratories on the relative antigenic composition of various strains and dissociants of Bacterium salmonicida. An experiment now in progress is a repetition and extension of provious work on the hypor- 4 immunization of trout by means of syringe inooulation. While this mothod is of course not applicable to hatchery procedures, it is intended to throw light upon the mechanism of protection in the cutthroat trout.

Diagnostic Services. A number of samples of diseased fish have been examined for tho Provincial Game Department. Cultures sent from California have again beon recognized as B. salmonicida, a furthor indication of the ondemic nature of this disease among west coast trout. - 92 -

D.C.B. Duff No. 69

References. (1) Rosenau, M.J. & J.F. Anderson - Hyg. Lab. Bull, No, 29, p. 67, I906. (2) Ratner, B. A H.L, Gruehl ~ J, Clin. Invest. 1£, 517, 1934.

(3) Walzer, M. - J. Imm. 14, 143, 1927.

(4) Burke, V, & J. Barnes - J. Infect. Dis. 39., 231, 1926. (5) Dow, R,P, - Can. Pub. H. Jour. j&, 570, 1940. (t>) Eraser, D,T,, E,L, Davey A K.C. Halpern - Can. Pub. H. Jour. ]&, 376, 1940.

(7) Bunyea,, H. - Jour. Agric. Ras, 34. £23* l$2Tl* ^ - 93 -

C, and E. Berkeley No. 70 Polychaeta of the Pacific Coast of Wp.rth America. We referred in last year's report to the collection of Poly chaeta from southern California made by Prof. MacGinitie of tho Corona Del Mar biological station, work which we had then completed. Since that time we received another collection made at Santa Cruz Island, in the same region, by Dr. W.G. Hewatt of the Texas Christian University. tfe, therefore, withheld publication until we could embody the results of a study of Dr. Hewatt's material with the rest. This study has added a few species to the list, some of which are of particular inter est. An account of the whole work is now in the hands of the Southern California Academy of Sciences, who will publish it shortly. We have now resumed work on the material collected during the cruise of the Canadian Hydrographical Survey boat "flfa. J. Stewart" on the west coast of Vancouver island in 1935. A comparison of the forms occurring m that locality with those obtained from California and from the east coast of Vancouver island is proving very informative. This work is more than half completed.

'Nf

-n - 94 -

J.A. Munro No. 71

a Studies of the Economic Relations of Waterfowl.

'» The life histories of waterfowl are involved to varying de- Q grees with those of other aniraals including commercially valuable fish esi Some species of birds feed upon the eggs and fry of Pacific salmon and trout and upon herring and herring eggs. Others compete with the Saimonidae for various organisms which enter into the diet of both (birds and fishes). Most of the bird species concerned are themselves of economic value. The object of this study is to obtain, through field work and by the examination of the stomach contents of birds, data which will be of value in connection with fish and game management and wildlife con servation generally. During the current year field studies were conducted in the Okanagan, I\Ticola, and Cariboo regions and on Vancouver island. Work in I the laboratory consisted of the examination of one hundred and forty-six j duck stomachs of twenty-one species, of eighteen coots, and of eleven ,• other water birds of seven species. The results are briefly summarized <§^„ in the following paragraph: 7 Horned Grebe - odonata nymphs, corixids, other aquatic insects in four, Gammarus liranaeus ninety per cent in fifth specimen; Pied-billed Grebe - sculpin, 65 per cent, Planorbis sp. 1 per cent, balance of con tents aquatic insects; Red-backed Sandpiper - microcaddis, mollusks, Kyallela azteca; Red Phalarope - corixids, seeds aquatic plants; Glaucous- winged Gull - barnacle scutes; Ancient Murrelet (sea bird collected in Okanagan) Gammarus limnaeus exclusively; Mallard - seeds Scirpus ameri- canus chief item, aquatic insects of next importance; Gadwall - Lemna *S trisulca, Zygnema, seeds Scirpus americanus; Baldpate - Elodea canadensis, ^mm^ seeds Scirpus americanus; Pintail - chironomid larvae (blood worms), ••Jlr*- tubers, seeds Potamogeton pectinatus; Green-winged Teal - aquatic in sects, seeds aquatic plants, mollusks, bulbils and oospores of Chara; Blue-winged Teal - amphipods, aquatic insects, mollusks, seeds and frag ments Potamogeton pusillus, seeds Scirpus americanus; Cinnamon Teal - corixids, Pianorbi3 sp., seeds aquatic plants, Shoveller - seeds'Rumex maritimus, Sparganium sp., Potamogeton pectinatus, corixids, Planorbis sp.; Redhead - seeds Myriophyllum spicatum, Ceratophyllum demersum, Scirpus americanus; Ring-necked Duck - stickleback, chironomid larvae and other aquatic insects, tubers, buds, seeds Potamogeton pectinatus, Pisidium sp.; Canvas Back - corixids, odonata nymphs, seeds Cerato phyllum demersum, Scirpus americanus; Greater Scaup Duck - chironomid larvae, corixids, odonata nymphs, Littorina scutulata, Pisidium sp., seeds Ceratophyllum, Scirpus, and Potamogetons; Lesser Scaup Duck - amphipods, gastropods, seeds aquatic plants including Myriophyllum spicatum, Ceratophyllum demersum, and Scirpus sp.; American Golden-eye - chironomid larvae, corixids, odonata nymphs, caddis, microcaddis, %[ leeches, seeds aquatic plants; Barrow's Golden-eye - amphipods, chiro y-~ nomid larvae^ caddis; Buffle-head - Hyalella azteca, odonata and mayfly nymphs, chironomid larvao, corixids, bryozoa, leeches, marine iaollusks, seeds aquatic plants; White-winged Scoter - Gammarus limnaeus, corixids, - 95 - i J .A. Munro No. 71

RuWT^3F^lmFhi™S od™»J °hironomid **"*», odonata nymphsT

americanus.«sVt^O^^^ plant ^^1, seeds chiefly Scirpus

.-^'aE