Control Was Made Easier by the Fact That Access to the Lake Is Usually Confined to One Place

Control was made easier by the fact that access to the lake is usually confined to one place. Fishing from the banks was not allowed, and fishing hours were defined as between 5:00 a.m. and just after sunset. The creel limit was ten fish per angler per day. Statistics were facilitated by the fact that boats were rented by the hour and those bringing their own boats were registered and charged a small fee.

BALDWIN POND

County: Stanstead Range: XI Township: Barnston Lot: 16

Longitude: 71°53' Latitude: 45°01' Altitude: 1, 575 feet

Area: 67 acres Maximum depth: 27 feet Average depth: 12. 5 feet

Legend:

.1.10-isobath in feet ::___:: road

0' 500' 1000' 1500' Scale:

Soundings: Calibrated metal sounding line with lead by Louis-Roch Seguin, August 4, 1949.

Figure i.-Description and location of Baldwin Pond. 31

There was a good season's fishing with 5,407 speckled trout reported. Distri- bution was as follows:

May 23 ...... 2,759 ( 384 anglers) June...... 818 May 24 ...... 351 July...... 515 End May ...... 500 August...... 253 September...... 211 Total for May...... 3,610

These figures show a great number of speckled trout taken during the season but tell nothing about the yield per hour per angler. As there are some gaps in the data as to the number of anglers and hours of fishing, particularly for July, we have omitted these doubtful periods from our compilation. In Table I we have a study of fishing yield.

Table I

Yield per angler in Baldwin Pond during 1955 season.

No. of Trout Date Fishing Trout per Trout per anglers caught hours angler hour

May 23 ...... 374 2,670 1,212 7.18 2.20

May 24 ...... 86 351 299 4.08 1.17

End May ...... 193 413 464 2.14 0.89

Total: May ...... 653 3,434 1,975 5.26 1.74

Total: June ...... 366 744 952 2.03 0.78

July 2,3,4 ...... 36 93 111 2.58 0.83

Total August ...... 144 234 360 1.62 0.65

Total September ...... 104 186 254 1.79 0.73

Total: Season ...... 1,303 4,691 3,652 3.60 1.28

The results of fishing in a lake during the first open season after it has been rehabilitated are highly significant. A yield of 1.28 trout per hour per angler through an entire season is very good fishing. Comparison is made with results in other Canadian and American lakes in Table II. In the light of these figures we can place Baldwin Pond in the category of excellent speckled trout fishing water. By continuing to gather angling results during future fishing seasons, biologists will be in a better position to know what trend to follow in future management work to give the anglers of Baldwin Pond continued good returns. 32

Table II

Comparative yields of various Canadian and U.S. speckled trout lakes.

Trout Creel census Lakes and Places Authors per hour duration

0.2-1.9 Crecy Lake, N.B ...... 1943-1953 Smith, M. W. 1955 0.3-1.2 Banff National Park (3 lakes), Alta. .. 1951 Cuerrier and Ward, 1952b 0.2-0.8 Nova Scotia (4 lakes) ...... 1945-1947 Smith, M. W. 1952 0.87 Michigan (trout waters) ...... 1928-1947 Fukano, K. G. 1948 0.8-1.2 Cape Breton Highland National Park 1951 Cuerrier and Ward, 1952a (2 lakes) 0.2-3.3 Jasper National Park (111akes), Alta.. 1951 Cuerrier and Ward, 1952b 1.1 Charlotte County, N.B. (8lakes).... 1941-1947 Smith, M. W. 1952 1.2 Gull Lake, California ...... 1940-1941 Curtis, B. 1951 1.28 Baldwin Pond, Quebec ...... 1955 Present document 1.7 Castle Lake, Calif ...... 1947-1949 Curtis, B. 1951 1.4-2.5 Montague Pond, P.E.I ...... 1943-1953 Smith, M. W. 1954 3.1-4.0 Fundy National Park, N.S. (3 lakes). . 1951 Cuerrier and Ward, 1952a

Fishing Methods On the opening day, May 23, anglers were asked about their method of fishing. As Table III shows, the majority of anglers went in for just one type of fishing. How- ever, a few tried two or even three methods. In these latter cases it was hard to find out just what device captured the fish, so we have refrained from detailing them. Still fishing was the most popular and had excellent results. However, although more trout were taken this way, spinning proved more efficient, although it was less used because of the more expensive equipment needed. On May 23, still fishers caught 2.04 speckled trout per hour while spinners took 2.95 speckled trout. The best spinning device proved to be the Mepps Spinner. We believe some fly fishermen became dis- couraged and tried another method. These were probably the less expert at the exacting sport of fly fishing, because those ten who adhered to fly fishing achieved the same fishing efficiency as those who confined themselves to the easier method of still fishing. Only twelve fishermen finished the day with empty creels.

Table III

Fishing methods used at Baldwin Pond May 23, 1955.

Fishing method No. of anglers Trout taken Trout per hour

Still fishing (worms) ...... 240 1,667 2.04

Spinning ...... 51 417 2.95

Spinning and still fishing ...... 45 374

Fty ...... 10 96 2.04 Fly and spinning ...... 3 24

Still fishing, spinning, By ...... 4 40 33 Growth of Speckled Trout In addition to registering the fishing results on May 23 and 24, the three different plantings of speckled trout were measured and weighed in order to study their growth in Baldwin Pond. As only legal-length (7 inch) fish were reported, it was necessary to await the opening of the 1955 season to get the results of the fall planting of 1954. However, there were a few clipped-fin specimens reported from the fall planting of 1954 and the spring planting of 1955. Table IV shows the minimum, average and maximum size per group. A glance at age groups, say yearlings, from the same hatchery but planted at different seasons shows that those planted as fingerlings in the fall and which consequently spent the winter and early spring in Baldwin Pond had better growth than those of the same age which wintered at the hatchery and were planted at the beginning of spring preceding the fishing. True, there are not many of these latter; but it does seem for these two days that the difference is remarkable.

Table IV Length and weight of speckled trout taken May 23 and 24, 1955 for different plantings.

Length in mm. Weight in oz. Group Age Date Number Min. Ave. Max. Min. Ave. Max.

Not marked ...... II May 23 2,673 155 245.5 310 1.00 4.75 9.00 May 24 317 180 239.7 285 1.75 4.32 7.00

Left pectoral fin off' ...... I May 23 52 165 206 . 7 240 1.25 2.88 5.00 May 24 17 170 204.7 244 1.50 2.73 6.00

Right pectoral fin off".... I May 23 4 175 194 . 5 221 1.75 2.37 3.00 May 24 4 182 193.0 200 1.75 2.18 2.50

' Planted in fall 1954 as fingerlings. Planted in spring 1955 as yearlings.

Speckled trout hatchery-reared a year and planted in Baldwin Pond reached, in length and weight, a year later, the following measurements: the 2,673 speckled trout of the same age measured from 155 and 310 mm. (6 to 12 inches) and weighed from 1 to 9 oz. (28 to 255 grams). Average length was 245.5 mm. (9.7 inches) and average weight 4.75 oz. (135 grams). For a population of speckled trout of the same age, say two years, it is highly interesting to know at that time of the year, in this case May 23, the trout's conditioning factor and relate this to the conditioning factors of other speckled trout in comparable circumstances. To this end we use the formula set up by Klak (1941), English system: CF - W X 100.000 L3 where CF stands for coefficient of condition in the English system, W stands for weight in pounds L stands for total length in inches 100,000 is a multiplier used to establish the decimal point at the right of two significant figures. 34

Taking the average length and weights found in Baldwin Pond, we find CF to be 32.5, with minimum 24.9 and maximum 38.1. Table V compares these results with those obtained by Greene (1952) from speckled trout of the same age in Stillwater Pond, Putnam County, New York.

Table V

Conditioning factor of speckled trout in Baldwin and Stillwater Ponds.

Maximum Pond Group Area Number Age (acres) depth Conditioning factor (feet)

Stillwater...... Hatchery..... 189 II 55 30 ave. C(TL.) 40.3 Wild...... 149 II - - ave. C(TL.) 31.5

Baldwin...... Hatchery..... 2,673 II 67 27 ave. C(TL.) 32.5

Table VI distributes the 2,673 speckled trout by length and weight groups, so we can plot a growth curve (Fig. 2). Since it is illegal to take speckled trout less than 7 inches long, it is understandable that the number of captures of shorter fish is not representative of the population. Hence, in plotting the curve for this population we have omitted groups less than 175 mm. in average length.

Total length is related to weight according to the formula: W = cLn where W is weight in grams, L is total length in millimeters, c and n are constants to be determined, being characteristics of the lake and the fish.

These constants can be calculated according to the following formulae, which are solutions of the initial equation resolved by the method of least squares:

2;Log W. E(Log L)2 - Log L. 2;(Log L. Log W) L og c = N. 2;(Log L)2 - ( 2;Log L)'

Log W - (N. Log c) and n = 2;Log L

Here N represents the number of length-classes utilized in the calculations.

The length-weight relationship for Baldwin Pond can now be concretely expressed by: W = 2.995 X 10-5 L2.1e1 35

Table VI

Length-weight relationship of speckled trout in Baldwin Pond.

Total length in mm. by classes of 10 mm. Number Weight in of fish grams

151-160 ...... 1 28.35 161-170 ...... 2 31.89 171-180 ...... 4 53.15 181-190 ...... 11 55.08 191-200 ...... 21 68.32 201-210 ...... 38 87.82 211-220...... 133 94.34 221-230 ...... 288 109.45 231-240 ...... 554 119.06 241-250 ...... 644 129.72 251-260 ...... 505 148.40 261-270 ...... 271 165.36 271-280 ...... 132 184.15 281-290 ...... 40 200.57 291-300 ...... 24 218.51 301-310 ...... 5 241.67

2,673

280

240

200

160

120

0 50 100 150 zoo 250 300 350 LENGTH IN MILLIMETERS

Figure 2.-Length-weight relationship of speckled trout two vears old in Baldwin Pond. (Points are based on lengths and weights, Table VI.) 36

WIROMIMMINFA=

Operation and installation for weighing and measuring fish for creel census at Baldwin Pond in spring of 1956. Richard L. Séguin, author, (left), his brother, also a biologist, Louis-Roch 8éguin (right) and a local fisherman.

Summary Baldwin Pond appears well managed to date. After the poisoning of the unwanted species of fish during fall 1953, the lake was ready to receive the speckled trout plantings of 1954 and 1955. The 1955 fishing season gave the first pertinent data necessary for the management of the lake. Successive creel censuses were the best tool to check management. The opening day, May 23, was very successful; 384 anglers caught 2,759 speckled trout. To prevent rapid depletion of the fish from the lake, in co,operation with the 37

organized anglers, a limit was set of ten fish per angler per day. The 5,407 speckled trout caught and reported from May 23 to the end of September prove that the plantings of hatchery stock gave good results. The yield of 1.28 trout per hour per angler for all the season proves the value of the lake as a speckled trout producer. The anglers made an outstanding contribution during this study of the lake. Still fishing, the preferred method of the majority, proved very efficacious. On May 23, 2.04 trout per hour were taken by this method. Spinning and fly fishing which were less in vogue were also effective. These more active methods need a little more equipment and are not, perhaps, for the inexperienced angler. To make a complete analysis of the fishing methods it would be necessary to study the baits and lures used by anglers. As Baldwin Pond seemed overpopulated at the opening of the season, it is understandable that the conditioning factor may be a little low. It did not prove practicable to learn the factor at the end of the fishing season, but it is a reasonable hypothesis, which later studies may confirm, that the conditioning factor improves in Baldwin Pond as overpopulation is reduced by fishing. By continuing the census year after year, it would be easier to approach the answer. Other data will help to compare growth curve. The constant n = 2.781 characteristic of the fish is a little lower than the optimum n = 3, but it is still very good.

Acknowlcdgmen*, s This study has been carried out as part of the freshwater management programme of the Quebec Biological Bureau, directed by Dr. Gustave Prévost. This work has been greatly aided by Louis-Roch Séguin, biologist and director of the Eastern Town- ships Fish Hatchery. It is a pleasure to acknowledge the co-operation of all anglers of Baldwin Pond during the creel census. Thanks are due also to Paul Bouchard, Larry Wilson and other associates in the Quebec Biological Bureau for helpful suggestions in the preparation of this manuscript.

Literature Cited

CUERRIER, J.•P. and J. C. WARD. 1952a. Game'^sh creel census. Analysis of creel census cards received from Eastern National Parks during the 1951 angling season. Canadian Wildlife Service, National Parks Branch. Dept. of Resources and Development. Ottawa, Canada. 14 pages. 1952b. Game fish creel census. Analysis of creel census cards received from Mountain National Parks during the 1951 angling season. Canadian Wildlife Service, National Parks Branch, Dept. of Resources and Development. Ottawa, Canada. 31 pages. CURTIS, BRIAN. 1951. Yield of hatchery trout in California lakes. California Fish and Game, 37 (2): 197-219. FUKANO, K. G. 1948. General creel census of fishing, 1947. Michigan Conservation, 17 (12): 8, 9, 15. GREENE, WILLARD C. 1952. Results from stocking brook trout of wild and hatchery strains at Stillwater Pond. Trans. Am. Fish. Soc., 81 (1951): 42-52. KLAK, GEORGE E. 1941. The condition of brook trout and rainbow trout from four eastern streams. Trans. Am. Fish Soc., Vol 70 (1940): 283-289. SMITH, M. W. 1952. Limnology and trout angling in Charlotte County lakes, New Brunswick. Jour. Fish. Res. Bd. Canada, 8 (6): 383-452. 1954. Annual crops of speckled trout from a Prince Edward Island pond. Progress Reports of the Atlantic Coast Stations, Fish. Res. Bd. Canada, No. 58: 21•23. 1955. Fertilization and predator control to improve trout angling in natural lakes. Jour. Fish. Res. Bd. Canada, 12 (2): 210-237.

Lea's Hydrostatic Tag on Brook Trout and Atlantic Salmon Smoks by

M. W. Smith

Fisheries Research Board of Canada, Biological Station, St. Andrews, N.B.

Einar Lea first demonstrated the hydrostatic tag at a meeting of the International Council for the Exploration of the Sea in 1948. Since that time this tag has been used by a number of investigators on several species of fish. Dannevig (1953) obtained better returns with hydrostatic than either disk or strap tags in his studies of Norwegian cod. However, of more interest to us is the use of the hydrostatic tag on salmonoids. Went (1951, 1953) and Went and Vickers (1953) employed Lea's tag to advantage in tracing the movements of adult Atlantic salmon along the Irish coasts. Vibert (1953) tagged salmon smolts in southern France. A few of these smolts were later recaptured by mackerel fishermen well to sea in the Atlantic. Jensen (1955) recovered as high as 44.8 per cent of rainbow trout tagged and liberated in Roskild Fjord, Denmark. In 1950 the hydrostatic tag was used on brook trout and smolts of the Atlantic salmon at Ellerslie Brook, Prince Edward Island. Since jaw tags had been and were currently being employed on brook trout at Ellerslie, a comparison of results from two types of tagging was possible. This article presents the results.

Lea's Hydrostatic Tag The hydrostatic tag is illustrated in Figure 1. It has been previously described (Anon., 1953; Rounsefeld and Everhart, 1953; Went, 1951). Briefly, the tag consists of a small celluloid cylinder, plugged at both ends, and containing written instructions to the finder. It is attached by a bridle of stainless steel to a wire pin which pierces the flesh of the fish, usually just anterior to the dorsal fin (Fig. 2). Nylon line has been used for attachment in some instances (Collyer, 1954; Dannevig, 1953; Jensen, 1955). The specific gravity of Lea's tag is somewhat less than that of water. The tags used at Ellerslie Brook were obtained from Mr. Lea. They were three centimetres long and four millimetres in diameter and coloured yellow with blue-tipped ends. Printed on the tag were the words "Canada", "Reward", and "cut ends, message inside", and a serial number.

Tagging Tagging was done in connection with a study of the movements and populations in Ellerslie Brook of the brook trout primarily and the Atlantic salmon secondarily (Smith, 1951). Fish are captured in a two-way fish trap (trap A) as they move between 39 40 salt water and Ellerslie Brook. A second similar trap (trap B), situated 650 yards upstream, captures fish as they move between the lower and upper sections of the brook. The traps have been described previously and their efficiency evaluated (Smith and Saunders, 1956).

R.w,emd r- A 9 £ ►

Figure I.-Leâ s hydrostatic tag.

Figure 2.-Wooden "boot" in which trout were held during tagging.

Fish were held in a wooden "boot" while the hydrostatic tag was attached (Fig. 2). A continuous flow of water was supplied through a hole at the head of the "boot". The forward end of the "boot" was tilted downward to insure that the head of the fish was entirely submerged. Three "boots" with holes of different diameters were used. It was necessary to have the hole sufficiently large so as not to interfere with movements of the opercula in respiration, yet small enough to retain the fish well. The fish remained quiet during the entire tagging operation. The following operations were involved in the tagging. The flesh of the fish, just anterior to the dorsal fin, was pierced with a number 18 hypodermic needle. A right-angle bend was made at one end of the pin. The straight end of the pin was threaded through one loop of the bridle and then through the hypodermic needle. The needle was withdrawn, the pin cut to the desired length and the other loop of the bridle slipped over the pin. Finally the straight end of the pin was bent, using two needle-nosed pliers, one to hold the pin rigidly, the other to bend it. From June 14 to July 4, 1950, hydrostatic tags were placed on 100 brook trout as they moved into Ellerslie Brook through trap A. These trout had an average fork length of 18.4 ± 0.62 cm. Similarly, in the period November 16 to 20, 1950, 77 trout 41

were tagged as they went downstream through trap B, and 23 others when they were taken going to salt water through trap A. The mean fork length of these two groups of trout were 16.0 ± 1.60 and 16.3 ± 2.87 cm. respectively. During April and May, 1951, hydrostatic tags were also applied to 493 salmon smolts as they ran to sea through trap A. These salmon had an average fork length of 13.6 ± 0.57 cm. From June 1 to July 27, 1950, 152 trout were jaw-tagged when they moved into the brook through trap A. During the month of November, 1950, jaw tags were also applied to 312 trout as they ran downstream through trap B and 261 as they went to salt water through trap A. This tagging was a continuation of a programme already under way at Ellerslie Brook whereby all trout moving through the traps, up- or downstream, were tagged. The jaw tag was of the strap variety, applied in a circular form around the mandible of the trout (Smith, 1957).

Recapture of Tagged Trout Records of recapture of tagged trout were obtained in a number of ways: (1) from the traps, (2) by an annual creel census maintained on the brook and estuary, (3) during population studies in the brook, and (4) by capture in commercial smelt nets in the estuary. Table I Recaptures of trout bearing hydrostatic tags.

Number recaptured and time at large Manner of recapture <1 week 1 2-4 weeks 1 <4 weeks

I--Of 100 trout tagged in A trap going upstream

Angled ...... 2

In traps ......

During population studies ......

In gill-nets ...... 0 0

II-Of 77 trout tagged in B trap going downstream

Angled ...... o o

In traps ...... 24 4

In gill-nets ...... 6 0

III--0f 23 trout tagged in A trap going to salt water

In traps ...... 0 2 1

In gill-nets ...... 2 0 0 42

Table II

Comparison of number of recaptures of trout bearing hydrostatic and jaw tags

Number Recaptures Number of recaptures Place tagged tagged after 100 days at large Number I Percentage

I-With hydrostatic tags

A trap up ...... 100 20 20 1

B trap down ...... 77 45 58 5

A trap down ...... 23 5 22 1

Totals ...... 200 70 35 7

II-With jaw tags

A trap up ...... 152 93 61 36

B trap down ...... 312 214 69 111

A trap down ...... 261 179 69 94

Totals ...... 725 486 67 241

Hydrostatic tags. Data on the recapture of trout bearing hydrostatic tags are summarized in Tables I and II. Seven tagged trout were recaptured twice. The first time was in trap A as they ran to salt water. However, only the final recapture is recorded in the tables. There were 20 recaptures from the 100 trout tagged as they entered the brook. The majority (13) of these recaptures was made by anglers within a month of the tagging dates. The longest period that any of this group of trout was at liberty between tagging and recapture was 150 days. Forty-five (58 per cent) of the 77 trout tagged in trap B as they moved downstream were recaptured. These trout were tagged in November when there is a general movement from the brook to salt water. Accordingly, it is not surprising that the greater number of recoveries were made at the mouth of the stream in trap A within a few days. Six individuals were subsequently taken in gill-nets in the estuary within a period of a week. Five from this group of 77 trout were at liberty 132 to 151 days between tagging and last recapture. Five recaptures were made from the 23 trout tagged in trap A when running to salt water. Two individuals were captured within a week in gill-nets. The other three were taken in trap A as they re-entered the brook, one after being in salt water for 220 days.

7 43

In summary, 35 per cent of the trout bearing hydrostatic tags were recaptured. However, the majority of these trout were at liberty for the short period of less than one month. Recovery of tagged trout after having run to salt water was poor. Out of 200 trout with hydrostatic tags, only seven upon recapture had been at large for more than 100 days. jaw tags. As shown in Table II, 67 per cent of the trout that were jaw-tagged as they moved up- and downstream through the traps were recaptured. There was little difference in the percentage recovery between those tagged going upstream and those that were running downstream and into salt water. In contrast to the returns with hydrostatic tags, a large proportion (50 per cent) of the jaw-tagged trout were at liberty over 100 days before recapture.

Recapture of Tagged Salmon There are no records that any of the 497 smolts bearing hydrostatic tags were recaptured. Twenty adult salmon entered Ellerslie Brook during the fall of 1953 and 17 in 1954. None of these fish showed any evidence of having been tagged. It is possible that tags could have been shed without leaving recognizable scars. However, careful examination of each salmon failed to reveal any. No capture of salmon with hydrostatic tags has been reported from the commercial salmon fisheries of the Gulf of St. Lawrence area. Two hydrostatic tags that had become free from the fish were found. One was picked up on the shore of Lennox Island in Malpeque Bay at a point about two miles from the mouth of Ellerslie Brook. The other was found on the beach at Percé, Quebec, a distance (shortest) of about 150 miles from Ellerslie. Since the tag is hydrostatic it would act in the manner of a drift bottle after being freed from the fish.

Summary It was observed that the bridle of the hydrostatic tag became fouled with aquatic vegetation, especially filamentous algae, in Ellerslie Brook and estuary. In the brook, the trout, including tagged individuals, sought hiding places when disturbed. In so doing it seems probable that snagging of the bridle would frequently occur. It was further observed that the hydrostatic tag and its bridle tended to ride to one or the other side of the fish. As a result, the flesh of the fish was abraded by the end of the bridle where twisted around the pin and by the sharp free end of the pin itself. These conditions would militate against the survival of the fish and the retention of the tag. The brightly-coloured hydrostatic tags were surprisingly conspicuous in the clear water of Ellerslie Brook. The bright colours were intentionally used to aid in recovery of the tags, yet their conspicuousness might attract undesirable attention. In this regard it was noted that trout struck at the hydrostatic tags carried by other individuals. 44

The fortuitous recovery of two unattached hydrostatic tags that had been placed on salmon smolts has been noted. Five additional unattached tags were found in Ellerslie Brook. These had been put on trout. However, the loss of these tags did not necessarily indicate death of the fish as shown by the capture of seven previously tagged trout with scars at the point of attachment of the tags. It is pertinent to note that jaw tags were shed by trout without apparent injury to the fish. Of 1,625 trout that had been jaw-tagged and subsequently recaptured in trap A during 1950, 70 (4.3 per cent) had lost the tag. (In applying both the hydrostatic and jaw tags the adipose fins were removed, which aided in spotting the previously tagged fish.) Although tried on only a few salmon smolts, it was found that the tendency of the hydrostatic tag and its bridle to slip sideways was largely prevented by placing two small celluloid disks on the pin or either side of the fish and inside the attachment of the bridle. The disks added some encumbrance to the tag. Attachment with nylon line may have lessened the difficulty. Jensen (1955) experienced more recaptures of rainbow trout when nylon was used to attach hydrostatic tags. On the other hand, Collyer (1954) found that attachment with nylon gave no better results than when wire was employed in tagging the yellowtail (Seriola dorsalis). The hydrostatic proved inferior to the jaw tag when used on trout at Ellerslie Brook. Fouling and snagging of the tag, more to be experienced in the confined quarters of Ellerslie Brook and estuary than in open waters, were important factors conditioning the results.

Acknowledgment The assistance of J. W. Saunders, C. R. Hayes and Cyril Williams is gratefully acknowledged. This paper is published with permission of The Fisheries Research Board of Canada.

Literature Cited

ANONYMOUS. 1953. A guide to fish marks used by members of the International Council for the Ex- ploration of the Sea. J. Conseil., 19(2): 241-289. COLLYER, R. D. 1954. Tagging experiments on the yellowtail, Seriola dorsalis (Gill). California Fish and Game, 40(3): 295-312. DANNEVIG, GUNNAR. 1953. Tagging experiments on cod, Lofoten 1947-1952: Some preliminary results. J. Conseil., 19(2): 195-203. JENSEN, I. BOHUS. 1955. Transplantation experiments with rainbow trout in bracltish water. Salmon and Trout Mag., No. 144, pp. 157-161. SMITH, M. W. 1951. The speckled trout fishery of Prince Edward Island. Canadian Fish Culturist, No. 11, pp. 1-6. 1957. Comparative survival and growth of tagged and untagged brook trout. Canadian Fish Culturist No. 20, pp. 1-6. SMITH, M. W. and J. W. SAUNDERS. 1956. Efficiency of year-round operation of trout counting fences on a small stream. Ibid., No. 18, pp. 6,9. VIBERT, R. 1953. Voyages maritimes des saumons et retour d la Riviere natale. Bull. francaise Pisci- culture, No. 170, pp. 5-23. WENT, A. E. J. 1951. Movements of salmon around Ireland. I. From Achill, Co. Mayo (1948 to 1950). Proc. Roy. Irish Acad., 54 (section B): 169-201. WENT, A. E. J. 1953. Movements of salmon around Ireland. III. From Carnlough, Co. Antrim (1950 and 1951). Ibid., 55 (section B): 209-223. WENT, A. E. J. and K. U. VICKERS. 1953. Movements of salmon around Ireland. V. From North Country Antrim (1951 and 1952). Ibid., 56 (section B): 13-28. EDMOND CLOUTIER, C.M.G., O.A., D.S.P. QUEENS PRINTER AND CONTROLLER OF STATIONERY OTTAWA, 1957. ISSUE TWENTY- ONE DECEMBER - - 1957

THE CANADIAN FISH CULTURIST

LIBRARY FISIiERIES AND OCEANS BIBLIOTHÈQUE PÉCHES ET OCPANS

Published at Ottawa by The Department of Fisheries of Canada CONTENTS Page The Importance of Size in the Change from Parr to Smolt in Atlantic Salmon-P. F. ELSON ...... 1 Using Hatchery-Reared Atlantic Salmon to Best Advantage- P. F. EI.SON ...... 7

Number of Salmon Needed to Maintain StocksP. F. ELSON 19 The Role of Hatcheries in Securing Maritime Stocks of Atlantic Salmon-P. F. EISON ...... 25

The Canadian Fish Culturist is published under the authority of the Minister by the Department of Fisheries of Canada as a means of providing a forum for free expression of opinion on Canadian fish culture. In the areas of fact and opinion alike, the responsibility for statements made in articles or letters reste entirely with the writers. Publication of any particular material does not necessarily imply that the Department shares the views expressed. In issuing The Canadian Fish Culturist the Department of Fisheries is acting only as an instrument for assisting in the circulation of information and opinion among people in the fish culture field. Those who may wish to discuss articles which have been published in The Canadian Fish Culturist are encouraged to do so and space will be made available.

Correspondence should be addressed to the DIRECTOR, INFORMATION AND EDUCA• TIONAL SERVICE, DEPARTMENT OF FISHERIES, OTTAWA, CANADA.

Published under Authority Of HON. J. ANGUS MACLEAN, M.P., Minister of Fisheries

I The Importance of Size in the Change from Parr to Smolt in Atlantic Salmon by P. F. Elson

Fisheries Research Board of Canada, Biological Station, St. Andrews, N.B.

The factors associated with the metamorphosis of young Atlantic salmon from river-dwelling parr to seaward-migrating smolts have not been very well defined. Size has been regarded as an important factor by some investigators and relegated to a minor position by others. Pyefinch (1955) gives a useful review of literature on the subject of this change from parr to smolt. Svardson (1955) suggested a set of hypotheses involving "physiological age" of the individual fish in combination with environmental considerations. Most investigators have concentrated their attention on the young salmon near or during the immediate season of migration. Vibert (1950), however, offered evidence that age at migration had been set by the end of the first winter of river life. Salmon production of the Miramichi River system, in New Brunswick, has been investigated by the Fisheries Research Board of Canada for the past seven years. Part of the program calls for knowledge of the numbers of smolts produced. Direct counts have not been attempted in the larger tributaries because log drives prevent weir operations during part of the run. Indirect estimates are consequently necessary. These rely upon a combination of seining and electrofishing done late in the summer. In the Miramichi system approximately one-third of the smolts migrate as two-year-olds and two-thirds as three-year-olds. Observations by Canadian investigators indicate that size, or physiological condition associated with size, of parr in their pre-smolt year seems to be a more important factor than age in the change from parr to smolt. As a working approximation an arbitrary dividing line between large parr, likely to become smolts next spring and small parr, not likely to, has been set at about 10 cm. or 4 in. total length, measured from tip of snout to tip of tail. From the facts assembled here it appears that there is justification for using this convenient rule of thumb. Evidence from hatchery yearlings In mid-June, 1945, hatchery-reared yearling Atlantic salmon of sea-run, and of lake, or landlocked, stocks were being planted in the Rawdon River, Grand (Schuben- acadie) Lake system of Nova Scotia. It was observed that the larger fish of both stocks had a heavy, silvery coating of guanin on their scales, much more than is normal for river-dwelling parr of comparable size. Other smolt characteristics such as low condition factor and easily loosened scales were apparent to a minor degree. Enquiry to the Department of Fisheries revealed that all the fish "had been fed (chopped) plucks and fish through January to the time of examination (June 16), except that they

1 96844-6-1 2

were fed fish only during the second and fifth weeks of May. It is quite possible that the thyroid gland was included in the plucks which were fed...... ". Hoar (1939) has shown that the change from parr to smolt stage, which he terms a true metamorphosis, is associated with greatly increased activity of the thyroid gland. Robertson (1949) showed that in large (9 to 10 inches) two-year-old rainbow trout (Salmo gairdneri) the silvery smolt stage could be induced by intra-muscular injection of mammalian thyroid extract. Hoar (personal communication) has expressed his opinion that "certainly fresh thyroid gland in the food supply should tend to initiate a smolt transformation". At the time of the observations above, approximately 300 fish from each stock were taken at random from the supplies in the hatchery's holding ponds. Each fish was classified according to the extent of its silvery colour as "parr", or "smolt". The fish were measured and mean lengths were calculated for each of the categories in both stocks. For both kinds of salmon there was a fairly clear-cut division into large, silvery "smolts" and into smaller "parr" without obvious silvery colour. These size groupings are shown in Figure 1. There was some overlapping of sizes between parr and smolts but there was no silvery coloration on fish of sea-run stock under about 112 cm. or on lake stock under about 16- cm. long. Comparison between sea-run and lake salmon is shown in Table I.

Table I.

Lengths (cm.) of yearling salmon of sea-run stock and of lake stock, from the Schubenacadie Grand Lake Rearing Ponds, Nova Scotia. (Measured June 16, 1945.)

Parr Smolts

Number Mean Standard Number Mean Standard Specimens Length Deviation Specimens Length Deviation

Sea-run ...... 264 10.2 +1.0 65 14.2 +1.3

Lake ...... 238 9.7 ±1.4 52 13.8 +1.5

The difference in size of the two stocks seems largely attributable to the lake salmon growing more slowly than the Atlantic salmon. In keeping with this a slightly smaller proportion of lake salmon (18%) than of sea-run salmon (20%) showed the change from parr to smolt. Wilder (1947) also found relatively slow growth to be a characteristic of lake salmon. The data show that both the sea-run and lake salmon stocks conformed to a general size characteristic in relation to the incidence of guanin deposition. The evidence does not show that the fish became smolts at a specific size of about 10 cm. Rather it shows that, for both stocks, only those fish which had already reached 10 cm. were subject to the change from parr to smolt. This evidence of a size prerequisite for transformation of hatchery-reared fish should be useful to fish culturists in planning salmon-rearing programs for particular purposes. For example, in a stream where smolts are typically three years old, fingerlings planted in one summer would be expected to produce a smolt run three years 3

120

100

20 tL O 1 5 LENGTH (CM)

LAKE SALMON YEARLINGS

SMOLT

.0 a' ^e ► .L: , I _s • 5 10 IS 20 LENGTH (CM)

Figure 1.-Length-frequency distribution of hatcheryreared yearling salmon of Atlantic and lake stocks, classified as parr haring no obvious silvery guanin deposit on the scaks, and smolts having a heavy guanin deposit. Only fish over Ili cm. (Atlantic) and over 10J cm. (lake) showed this smolt characteristic. (Fish examined at Grand Lake Rearing Ponds, N.S., June 16, 1945) later; planting small parr should yield a run only two years after planting, and planting large parr one year after. Liberating smolts directly from the hatchery is proving worth while for one Swedish river from which salmon have been excluded by a dam (Carlin, 1955). Programmes using hatchery-reared smolts might benefit by consideration of such a size prerequisite. This subject is now being examined through experimental plantings. Evidence from Pollett River Smolts During the course of experiments involving Atlantic salmon smolt production from the Pollett River, Petitcodiac River system, New Brunswick, scales from several thousand smolts were examined. For the most part age determinations only were made. Here, 90 to 95 per cent of the smolts run as two-year-olds, and almost all the 96844-6--l i 4

rest as three-year-olds. Measurements of scale radii for the several years of river life were made on 250 typical samples from the 1950 smolt run. For the 14 three- year-old smolts measured, the calculated mean length and standard deviation at the end of the second winter was 10.7 ± 0.8 cm. For a representative group of 22 two- year-old smolts from the same migration, calculated mean length at the end of the second winter was 14.0 ± 1.7 cm. Examination of scales from Pollett parr collected monthly through most of a year indicated that usually about one-quarter of the growth for the year occurred during the laying down of the winter band, between about Mid- August and the following May. In consequence most fish which were about 11 cm. long by the time the winter band was completed were probably well under 10 cm. long in the preceding August: but those which were around 14 cm. long, or longer, in May must have been at least 11 or 12 cm. long in the preceding August. These Pollett young salmon conformed fairly well to the hypothesis that those which exceed a length of 10 cm. by late summer, become smolts in the following spring; while those which do not attain this size remain in the river for an additional year or more.

Evidence from Scotland-Thurso River Allen (1944) made an extensive study of salmon smolts from the Thurso River system of northeastern Scotland. In his Table VII he gives the lengths at smolt stage and calculated lengths at the end of the immediately preceding winter for 1,317 two-year-smolts. These are expressed as mean lengths for fish captured during successive five-day periods. The general mean for all fish at the end of the preceding winter was 10.6 cm. Only two groups totalling 50 fish, or 4 per cent of the collection, had a mean length under 10 cm.; and it was 9.5 cm. at the end of the preceding winter. Unfortunately Allen does not give first- and second-winter lengths for the small percentage of three-year smolts he recorded. However, the fact that 96 per cent of his two-year smolts reached the 10 cm. length at this time, and the rest nearly attained it, gives support to the hypothesis that there was a size prerequisite for the change, which was of the same general order as that observed in Canada. Allen concluded that "there is no critical length at which migration takes place". But he concurred with earlier authors, that "parr must attain some physiological condition associated, at least as an index, with a minimum size before the smolt migration takes place". The hypothesis that this condition is associated with a length requirement of about 10 cm. during the latter part of the pre-smolt year of growth, is not at variance with the conclusions drawn by Allen.

Evidence from England-The Cheshire Dee Jones (1949-Appendix, Table II) gives the calculated yearly lengths of 335 smolts taken from the River Dee, Cheshire, in 1938 and 1939. He had one-, two- and three- year smolts. Nearly 92 per cent of these fish satisfy the present hypothesis. Dee smolts which did not conform were those migrating at 1+ years. But even these had all made rapid growth during the first season-72 to 9z cm. against 4â to 5 cm. Rapid spring growth in the season of descent brought them to about 1212 CM- Since a large majority of the fish conform, the 10-cm. hypothesis should be useful in studying late summer parr populations in such a stream. 5 Evidence from France-River Adour Vibert (1950) gives calculated mean lengths at the end of each year of river-life for 492 smolts and 1,022 virgin salmon, in his Table XIV. Nearly all the fish conformed in that they did not migrate if they had not reached the 10-cm. length, but migrated in the next spring, after reaching or exceeding an approximate 10-cm. length towards the end of a growing season. One three-year smolt reached a length of 11 cm. at the end of its second year, as did one of Jones' three-year smolts from the River Dee. Since the 10-cm. hypothesis is based on mean lengths in late summer rather than the length of individual fish at the end of a year's growth these do not constitute exceptions. Among Vibert's virgin salmon, a group of 11 which had been three- year smolts had a calculated mean length of 13.6 cm. at the end of their second year of river life but still spent one additional year in the river. The fact that 99 per cent of his fish conformed to the 10-cm. hypothesis indicates that it would be a useful working rule here also. Discussion and Summary Evidence is presented from both sides of the Atlantic to show that, as a general rule, parr which have reached or exceeded a certain size towards the end of one growing season are likely to become smolts at the next season of smolt descent. A similar size prerequisite also appears to hold for a slower-growing Canadian lake salmon stock. In fact, the concept of a size prerequisite appears to have more fundamental significance with respect to change from parr to smolt than either age or size at the actual time of transformation. The hypothesis definitely does not specify that young Atlantic salmon become smolts as soon as they reach a length of 10 cm., but rather that they transform after reaching this length in the smolt-running season immediately following. Growth of parr is variable. Those which have nearly but not quite reached the 10-cm. length at the end of a summer will, if they grow fast, reach a much greater length, say 14 to 16 cm., at the end of the following summer, and make big smolts. Those which have just barely attained the required 10 cm. size at the right time will, if slow-growing, make little smolts, perhaps only about 12 cm. long. The 10-cm. hypothesis thus takes account of the extremely variable size-range of smolts from,many types of rivers. Some individuals destined to become older smolts may exceed this length, but seldom by more than two or three cm., at the end of their second last season of river growth. A few fish may not quite reach this size but still become smolts during the following spring if they grow fast and exceed the 10-cm. length during the smolt- running season. While not precise for individual fish, this 10-cm. hypothesis provides a useful and justifiable rule of thumb for gauging annual smolt production from streams. Canadian investigators of Atlantic salmon at present classify young salmon as "small" parr, those which will probably spend one or more additional years in river growth, and "large" or "pre-smolt" parr, those which are likely to descend as smolts during the next seaward migration. The arbitrary division between the two groups of parr has been chosen at a total length of 10 cm. or 4 in. reached during the latter part of a year's growth. Use of the hypothesis should contribute to effective planning of salmon-rearing programs for hatcheries, if special purposes can be served by planting larger fish. 6

Acknowledgments The Grand Lake salmon were examined while assisting Dr. A. G. Huntsman in his investigations there; he kindly made time available for the observations on hatchery stocks. Mr. William Cameron, Superintendent of the Grand Lake Rearing Ponds, assisted in handling these yearling salmon. Mr. W. G. Carson of this Station made the scale measurements for Pollett smolts while reading some 2,000 smolt scale samples from this stream. I thank these gentlemen for their help.

Literature Cited

ALLEN, K. RADWAY. 1944. Studies on the biology of the early stages of the salmon (Salmo salar). 4. The smolt migration in the Thurso River in 1938. J. Anim. Ecol., Vol. 13, pp. 63-85. CARLIN, BORJE. 1955. Tagging of salmon smolts in the River Lagan. Institute of Freshwater Research, Drottningholm. Report No. 36, Ann. Rept. for the year 1954 and short papers, pp. 57-74. HOAR, WILLIAM S. 1939. The thyroid gland of the Atlantic salmon. J. Morph., Vol. 65, No. 2, pp. 257-295. JONES, J. W. 1949. Studies of the scales of young salmon, Salmo salar L., in relation to growth, migration and spawning. Ministry of Agriculture and Fisheries, U.K. Fishery Investigations. Series I, Vol. V. No. 1, pp. 1-23. PYEFINCH, K. A. 1955. A review of the Literature on the Biology of the Atlantic Salmon (Salmo salar Linn.). Scottish Home Department. Freshwater and Salmon Fisheries Research. No. 9, pp. 1-24. ROBERTSON, P. H. 1949. Production of the silvery smolt stage in rainbow trout by intramuscular injection of mammalian thyroid extract and thyrotropic hormone. J. exp. Zool., Vol. 110, No. 3, pp. 337-355. SVARDSON, GUNNAR. 1955. Salmon stock fluctuations in the Baltic Sea. Institute of Freshwater Research, Drottningholm. Report No. 36, Ann. Rept. for the year 1954 and short papers, pp. 226-262.

VIBERT, R. 1950. Recherches sur le saumon de l'Adour (Salmo salar Linné). (Ages, Croissance, Cycle génétique, Races) 1942-1948. With an English summary. Annales de la Station centrale d'Hydrobiologie appliqué, Tome 3, pp. 27-149.

WILDER, D. G. 1947. A comparative study of the Atlantic salmon, Salmo salar Linnaeus, and the lake salmon, Salmo salar sebago (Girard). Canadian J. Res. Section D. 25, pp. 175-189. Using Hatchery-Reared Atlantic Salmon to Best Advantage by P. F. Elson

Fisheries Research Board of Canada, Biological Station, St. Andrews, N.B.

Good management of Atlantic salmon requires knowledge of how to get the best production of seaward-migrating smolts. This applies whether the stocking be through natural means or through the introduction of hatchery-reared stock. The Fisheries Research Board of Canada has, for a number of years, been investigating factors associated with the production of smolts from hatchery-reared stock. This is a preliminary survey of the results which can be found in condensed form at the end of the paper. Much of the following discussion will refer to numbers which can be produced, must be planted and so on. To use total numbers from experiments would have no significance for other areas than the one in which studies were made. Hence, important figures have been converted to easily imagined units. For general use the average number per 100 square yards has been selected-an area 30 feet by 30 feet, or about the size of an average house. Sometimes it is more convenient to think in terms of the length and width of a stream. For this a unit of one mile of stream 10 yards wide (176 of the square units) has been selected. The 100 square yard unit will be meant unless otherwise stated.

1. PREDATION BY MERGANSERS SHOULD BE LIMITED Several years ago White (1939) found that a stream stocked with native salmon gave more smolts when mergansers (Mergus merganser americanus) and kingfishers (Megaceryle alcyon alcyon) were kept away. A study has recently been completed on the effect of similar bird control in increasing the smolt yield from hatchery-reared fish. Experiments were done on a 10-mile stretch of the Pollett River, a tributary of the Peditcodiac system in southeastern New Brunswick. Table I shows the total number of two- and three-year-old smolts produced from seven different plantings of underyearling salmon. Two of the plantings were quite light; five were fairly heavy. However, as will be shown in Section 2, below, even the light plantings should have experienced a better survival rate than they did. If judged solely on a comparison of the results from the heavy plantings, there was an increase of about four times with birds kept away. If judged on the average "before and after" figures, by nearly seven times. Table I also shows the estimated number of parr in the river, midway between planting and smolt descent. It is on these parr that the birds do most of their feeding.

7 8

Table I

Smolt production from known plantings of hatchery-reared underyearling salmon in the 10-mile experimental area (approximately 435,000 sq. yd.) of the Pollett River, N.B.

Number of Resulting parr Year of Resulting underyearlings 1 year after planting 2- and 3-year planted planting smolts (nearest 5,000) (nearest 5,000) (nearest 1,000)

With no bird control 1942 15,000 under 5,000 2,ÔOo 1943 15,000 under 5,000 1945 1,000 250,000 10,000 5,000 Average ...... over 15,000 5,000 3,000

With control of inergansers 1947 275,000 25,000 22,000 1948 235,000 and kingfishers 45,000 14,000 1949 245,000 40,000 1950 19,000 245,000 55,000 24,000 Average ...... 250,000 40,000 20,000

Records were kept of all birds seen. The food recently eaten by many of those shot was studied. In Table II estimates are given of the number of fish which the birds could have eaten had there been no control. There are also estimates of the numbers that could have been eaten by those birds which escaped the patrols. Without control, enough mergansers visited the area to eat more parr than the river could support. But with control the remaining mergansers could take only about one quarter of the parr when there was a good parr crop. This is not enough to seriously affect the smolt-producing capabilities of the stream.

Table II Estimated numbers of fish which visiting mergansers and kingfishers could eat annually while on the 10-mile experimental area of the Pollett River.

Mergansers Kingfishers All fish Young salmon All fish Young salmon

Without Control ...... 180,000 75,000 120,000 10,000 With control ...... 20,000 10,000 10,0c0 1,000

Year-to-year records showed that when parr were more abundant, more mergansers came to the area. They came any time of the year that there was open water on the river (Fig. 1).

In general, the rate of production when mergansers had free reign was about one smolt per 100 square yards of stream. With bird control, production jumped to a rate of five to six smolts per 100 square yards. 9

Figure 1.-Mergansers on a New Brunswick salmon stream in mid-winter.

Surveys have shown that mergansers are as abundant on many Maritime streams as they were on the Pollett. But they are furtive birds and their abundance can easily be underestimated. To get the best results from hatchery-reared stock, or native stock too, for that matter, mergansers should not be more abundant than about one bird for every 10 to 15 miles length and 10 yards width of stream. Kingfishers, as shown in Table II, could not remove enough young salmon to seriously affect the smolt output unless the birds were to increase greatly in numbers. Their habits make them conspicuous along a stream. Because of this, in contrast to mergansers, it is very easy to overestimat^ their abundance and their importance to salmon. It is doubtful that the results of controlling kingfishers along most salmon streams would warrant the trouble and expense. However, if they were more abundant than two for every mile length and 10 yards width of stream, control might be worth while.

2. HOW MANY FISH SHOULD BE PLANTED?

To know how many fish to plant it is first necessary to know how many smolts an area can produce.

Smolt producing capacity of streams. The bird control studies mentioned above showed that the 10-mile stretch of the Pollett could produce smolts at a rate of about one per 100 square yards without bird control; but five with control. This was from 96844-6-2 10

planting fingerlings at rates of 55 to 60 per 100 square yards. To find out whether planting more fingerlings would give more smolts a planting was made at the rate of 213 fingerlings per unit area-about four times as many. Data are given in the two top rows of Table III. Smolts were produced at the rate of six per unit area. This is not sufficiently different from five smolts to be regarded as anything more than the difference between years with favourable or unfavourable conditions for survival. That is to say, the maximum capacity with bird control can be set at five to six smolts per 100 square yards. Table III

Smolt production and survival rates from heavy, medium and light plantings of underyearlings and one small accidental spawning in a 10-mile stretch of the Pollett River, N.B. Total experimental area = 435,000 sq. yd. (a) Accidental spawning resulted when exceptionally high freshet brought a few adult salmon past a closed fishway into the experimental area. (b) "Best survival rate" from planting to parr was approximately 25%. (c) "Best survival rate" from parr to smolts averaged approximately 65%.

Number of Survival rate Survival underyearlings planted One-year parr underyearlings Two•year smolts rate parr per 100 sq. yd. to parr % to smolts °J'o total per 100 sq. yd. per 100 sq. yd. total no.

925,000 213 16 8 6 23,750 34 240,000 55 15 27(b) 5 22,852 35 65,000 15 4 27(b) 2 8,052 51(c) accidental(a) over 1 under 1 3,556 80(c)

Observations on young salmon in Miramichi streams (Elson, 1957a) indicate that the smolt producing capacity of these streams, both without and with bird control, is similar to that found for the Pollett. A relatively infertile stream in the Gaspé peninsula of Quebec, the Port Daniel River, also gives smolts at the approximate rate of one per 100 square yards; not enough mergansers visit this stream to warrant their control.

Survival rates from fingerlings to smolts. The plantings which gave maximum smolt production may have wasted quite a few fry. In fact, compared with results of the planting made at a rate of 55 fry, planting 213 fry did waste about three quarters of the fish. To discover the "best survival rate" when no such wastage was involved, a planting was made at about one quarter of the first rate (actually, 15 fry per 100 square yards). This and other measurements of "best survival rate" from sparse populations, which means no unnecessary wastage, are given in the last two lines of Table III. The best survival rates from fry to yearlings were a bit over 25 per cent. The best survival rates from sparse parr populations to smolts averaged 65 per cent. This gives over 16 per cent as the survival rate from fry to smolts. To play safe the general rate from late summer fingerlings to two-year smolts should be 15 per cent. Although the smolt capacity of Miramichi streams is apparenly about the same as the Pollett, there is one important difference. Where nine out of 10 Pollett smolts are two years old, seven out of 10 Miramichi smolts are three-yearrolds. This means Miramichi part must go through an extra year in the river, during which there is some il

mortality. Here again the observations in the Miramichi area indicate that the survival rate over this extra year is of the same general order as the rate over the last year on the Pollett, i.e., about 65 per cent. So the survival rate for three-year smolts from late summer plantings of fry should be set at only 10 per cent. It would be useful to know from which streams to expect three-year smolts and which streams two-year-olds. Smolt age appears to be clearly associated with water temperature. In general, warm streams, say over about 68°F. (20°C.) in summer give two-year smolts. Cool streams, say 60°F. to 65°F. give three-year smolts. In still colder, or in very infertile streams, smolt age may be increased to four, five or even eight years; but smolts this old are uncommon in the Maritimes. Number to plant. The best number to plant is that which will give the full capacity of smolts, while giving the highest survival rate from the planted fish. Approximate figures for the largest numbers of fingerlings it would be worth while to plant are: In fertile streams with mergansers less than one per 10 miles x 10 yd. of stream. for two-year smolts 7,000 fingerlings per mile x 10 yd. wide for three-year smolts 10,000 fingerlings per mile x 10 yd. wide In barren streams, or fertile streams with mergansers more abundant. for two-year smolts 2,500 fingerlings per mile x 10 yd. wide for three-year smolts 3,500 fingerlings per mile x 10 yd. wide These are the highest figures. In most years planting 1,000 or so less should give nearly all the smolts the stream could raise. But planting many more is almost certain to result in waste of all the excess either through predation or starvation.

3. AMOUNT OF DISPERSAL AT PLANTING.

There are two aspects to this: the amount of scattering at chosen planting sites; and the distance between planting sites.

Dispersal at planting sites. The best amount of dispersal at planting received study on a section of the Pollett 10 miles above the area used for the studies of smolt production already mentioned. Three different degrees of dispersal were tested at each of three similar planting sites located two miles apart. The fish planted in each area were distinctively marked by removal of certain fins. Three years was required to complete the planting schedule. The results were measured by counting the smolts from each planting as they passed through the smolt trap 15 miles downriver (Table IV). The final results indicate that the different degrees of dispersal had no appreciable effect on the number of smolts produced. The difference in the various smolt r uns (Table IV) proves on analysis to be largely the result of good and bad years. Further, the bad years were years when mergansers were relatively abundant. No bird control, other than removal of merganser broods, was employed for these dispersal experiments. The important conclusion from this study is that nothing would appear to be gained 96841-6-2+} 12

Table IV

Smolts produced from plantings of 4,000 underyearling salmon given different amounts of dispersal at liberation.

Number of smolts produced Dispersal at planting, as nos. per 100 sq. yd. from 1948 from 1949 from 1950 from all 3 plantings plantings plantings plantings

50 ...... 1,183 1,911 618 3,712 500 ...... 1,329 1,407 868 3,604 5,000 ...... 767 1793 972 3,532 Total smolts from plantings of 1 year ...... 3,279 5,111 2,458 10,848

by scattering underyearlings, at planting, more widely than at a rate of 5,000 per 100 square yards at each planting site. A similar study of dispersal rates for yearlings is just being started.

Distance between planting sites. The natural spreading of the young fish away from their points of liberation was followed in several of the planting experiments mentioned above. The general picture was that the surviving fish spread fairly uniformly over the stream for about one-half mile up- and one mile downstream in one to two months' time. A year after planting they were fairly well scattered over one mile up- and downstream. A few fish travelled several miles farther than this. But the intervening stream was clearly not being fully utilized. Such lack of utilization did apply to an area two miles below the lowest liberation site through all the years of extensive planting for bird control and smolt capacity studies. It is therefore concluded that to get full utilization of a stream's resources planting sites should, ideally, not be over two miles apart. However, access for hatchery trucks will often make this impractical; in that event some, if not full use will be made of intervening areas, even if they be several miles long.

4. SIZE OF FISH TO PLANT.

Until recently most hatchery plantings of Atlantic salmon in Canada have liberated fingerlings around 12 -2 inches long. By holding the fish a year they can be grown to 3-6 inches long. Rearing large fish is more costly than rearing underyearlings. When is the extra expense warranted?

A common reason for planting larger fish is the hope that they may be able to escape enemies better than small fish. Consider, however, streams where mergansers are the chief enemy. Mergansers usually take parr from about three inches upwards, but seldom take two-inch fish if larger are available. In such streams newly-planted parr may not be competent in avoiding their new enemies. But their brothers planted as much smaller fingerlings are well adjusted to life in streams by the time they reach this size. Newly-planted fish face not only predators, but also competition with native fish of their own kind. As shown by Miller (1953) for a related species (Salmo 13

clarki) of the west coast this and the strain of adjustment to a new habitat may reduce survival to as low as 0 per cent-3 per cent, even for fish larger than any salmon parr. The results of some small-scale experiments on yearlings planted in two streams of the Petitcodiac River system in New Brunswick are informative. Holmes Brook and Bennett Brook are each about five miles long and are from two to five yards wide. Both have small populations of native salmon (about two per 100 sq. yd.). In the Holmes Brook more than half the native salmon migrated as three-year smolts, the rest as two-year-olds. Practically all Bennett smolts were two-year-olds. Other fish present in both streams were eels, homed dace (Semotilus atromaculatus), Lake North- ern chub (Couesius plumbeus), black-nosed dace (Rhinichthys atratulus) and brook trout (Salvelinus fontinalis) which are fished by local anglers. Occasional kingfishers visited the streams, but mergansers seldom if ever. Thus, the two brooks offered conditions similar to those on many Maritime salmon streams when mergansers are not abundant. The hatchery yearlings, averaging about 10 cm. (4 in.) long, measured from tip of snout to end of tail, were divided into a number of small groups, some of which were of ordinary sea-run stock and some lake stock. Each lot was distinctively fin- clipped before planting. No differences in behaviour or survival value of the two stocks was noted. But considering the differently-marked groups as comprising different experiments adds to the value of the final result. All those surviving to smolt stage migrated the year after planting, as two-year-olds. This applied in both brooks. The results are given in Table V. They may be summed up, in the

Table V

Smolt production and survival rates from July plantings of hatchery-reared one-year-old Atlantic salmon parr in tributaries of the Petitcodiac River, N.B. (In 1945 both Atlantic and Lake salmon were planted in about equal numbers: in 1946 only Atlantic salmon.)

Date and place of planting.

Year Brook Miles above Number Number of Survival smolt trap planted smolts rate

1945 Holmes ...... 26 (Atlantic) 12 26 1 23 26 (Lake) 13 25 26 « 1945 Bennett ...... 24 (Atlantic) 26 26 31 (Lake) 25 " 26 " 1946 Bennett ...... 50 (Atlantic) 150 117 53 50

Total 682 1-year parr 90 2-year smolts

Mean survival rate and its standard error ...... 13 ± 3(*^ 14

statement that it is unlikely that similar plantings of yearling parr would experience an average survival rate of less than 5 per cent or more than 20 per cent.

This is not materially better than the 15 per cent survival rate which can be obtained from underyearlings (see Section 2). But the yearlings yielded smolts one and two years earlier than underyearlings would have. On the basis of these data we must conclude that if the object of planting young salmon is merely to obtain the greatest number of smolts, then proper numbers of underyearlings should be planted. Whether a similar result would apply if such predacious fish as white perch, yellow perch, and fallfish (Semotilus corporalis) were abundant, as they are in some salmon streams, has not been investigated. Another conclusion is that if the object is to obtain smolts as soon as possible, then larger parr should be planted. By properly selecting the fish it should be possible, in a three-year smolt stream to reinforce smolt runs occurring any time between planting and three-years later as desired. In order to become smolts, parr must, as a rule, reach a length of about 10 cm. (4 in.) some months before the spring period of smolt migration (Elson, 1957b). Thus, as happened in Holmes Brook, yearling fish which will reach this length in the year of planting will make two-year smolts the next spring, even though the normal smolt age for the stream is three years. Fish which will reach only eight or nine cm. in the year of planting should make smolts two seasons later.

Hatcheries rear some young salmon to smolt stage as yearlings. Will they migrate to sea immediately, if liberated? Hoar (1939) found that the change to smolt stage was associated with increased activity of the thyroid gland. But he also found that if smolts were forced to remain in fresh water after the normal migration period this gland tended to revert to the parr condition. The migratory urge may be lost. This is indicated by the behaviour of some of the yearling fish planted in Bennett Brook in July, 1946. Among the 420 fish liberated, 16 per cent appeared to be smolts, similar to the condition Elson (1957b) reported for other hatchery yearlings. None of these Bennett fish was taken in the smolt counting weir below, in the year of planting. Several were recognized living near their original planting sites, during the summer. Occasional post-smolts which failed to get to sea have been observed in other places. They appeared to be slowly reverting to characteristic parr appearance and behaviour. Hence it would appear advisable to liberate hatchery-reared smolts early in the spring -say in May-if advantage is to be taken of their migratory urge.

Occasions may occur when the expense of planting larger fish than normal underyearlings would be warranted. The recent decimation of Miramichi young salmon by spraying of DDT has been reported by Kerswill and Elson (1955). Plantings of f assorted sizes of young salmon could properly reinforce the various year-classes. But these authors found that more large fish survived the spraying than small. It may be necessary that planted fish be subjected to spraying in order to preserve the forests. If so, it would seem advisable to plant them at the largest size reasonably possible. Spraying gives rise to one situation where only large fish could remedy some of the deficiencies. 15

5. TYPES OF BOTTOM FOR PLANTING. In the experimental plantings described, most fingerlings were liberated in water less than one foot deep, usually in rapids. Studies of the distribution of native young salmon have shown that this is the kind of habitat they prefer. It would be incurring unnecessary risk to liberate such small fish, not yet adapted to the rigours of life in streams, in pools where larger trout or other fish could readily gobble them up. Equally hazardous are areas with large cobble and rocks where eels are abundant. In the Pollett plantings, eels from about a foot upwards in length, were found to take many fingerlings within a few hours of planting in such areas.

6. EELS AS A LIMITING FACTOR FOR SURVIVAL OF UNDERYEARLINGS. Various observations made during earlier studies, as well as the Pollett experiments, have led to the belief that eels may sometimes limit survival between fry and parr stages. The dispersal experiments described in Section 2 were carried out above the 15-foot drop of Gordon Falls. The falls is a partial barrier to eels, there being one-third as many of the sizes that eat salmon above as compared to below. Up here over 45 per cent survival to smolt stage was reached in one case, the average for the nine plantings being 30 per cent (Table III). On the lower, main experimental area, with an abundant eel population, the comparable survival rate to smolt stage was only 15 per cent. Note that this concerns survival rates from sparse populations. The bird control experiment showed that mergansers act as a limiting factor atter eels have done their damage. Hence control of the birds is of first importance. Studies of eels in several New Brunswick streams indicate that they are abundant in streams flowing into the Bay of Fundy, e.g. the Pollett and several tributaries of the St John (Godfrey, 1956) but that in Gulf of St. Lawrence drainages, as represented by the Miramichi River, also Ellerslie Brook in P.E.I., eel populations in the salmon- growing reaches are relatively unimportant. Higher yields of salmon, because of better survival rates, might be expected from both hatchery and native stocks where eels are scarcer, providing predation on older parr and smolts, as by mergansers, is not the limiting factor. So far no practical way of controlling eels has been devised.

7. VALUE OF ADULTS DERIVED FROM HATCHERY STOCK. Returns to the Pollett River of mature salmon derived from stocks collected in other rivers but planted as underyearlings in the Pollett have been almost negligible. These planted fish have contributed to distant commercial fisheries; but were not identified in any river where close inspection of the stock was maintained. At present a test is under way involving the returns from 800 marked smolts of Pollett stock, reared f to the usual underyearling stage in a hatchery, then planted back in the Pollett. Summary This summary is given in the form of brief, if somewhat categorical statements, for convenient use of the practising salmon-culturist. The statements should certainly not be regarded as the "final word" on subjects which are still under investigation. But they do have more background of observed facts than has been generally available. 16

(1) Control of mergansers. Best production of smolts from plantings of hatchery- reared Atlantic salmon, or from native salmon should only be expected if mergansers do not exceed an abundance of one bird for every 10 to 15 miles and 10 yards width of stream. Control of kingfishers is of minor importance compared to control of mergansers; kingfishers may be tolerated up to, but not above, an abundance of two per mile x 10 yards width of stream. q

(2) Numbers to plant. For streams on which mergansers will not greatly exceed the figure of one per 10 miles X 10 yards width, and which flow through geologically fertile areas the maximum number of smolts likely to be produced is around 1,000 per mile of stream 10 yards wide. To get this number of smolts about 7,000 underyearlings should be planted, per similar area, in streams with mostly two-year smolts and about 10,000 in streams with mostly three-year smolts. For streams with no bird control or those in geologically infertile areas, e.g. flowing largely through granite country, these figures should be reduced to about 2,500 and 3,500 fish for a maximum expectancy of 400 smolts.

(3) Amount of dispersal at planting. When liberating underyearling salmon at planting sites, nothing is gained by scattering them more widely than at a rate of 5,000 fish in an area of 100 square yards. Given this start, as many will survive as if they had been dispersed more widely. Planted underyearlings will move, in good numbers, about one mile up- and downstream from planting stations. So stations should ideally be about two miles apart. But some fish will disperse to as much as several miles away, so that partial use of intervening water will be made even if stations must be farther apart.

(4) Size of fish to plant. Where the object is simply to obtain smolts, underyearlings will give as good results as larger parr. But by using appropriate sizes of parr, smolt runs can be produced earlier than with underyearlings. For example, in a typically three-year smolt stream: (a) underyearlings will yield smolts three years after planting: (b) parr which will reach a length of eight or nine cm. in the year of planting should produce smolts two years after: (c) parr which will exceed a length of 10 cm. in the year of planting should produce smolts one year after: (d) smolts planted from early spring until late in May might well migrate to sea shortly after being liberated but smolts planted in streams after June seem unlikely to descend in numbers until the following year. Since underyearlings are very susceptible to DDT poisining and larger parr less so, the size of fish to be used for rehabilitating streams in DDT-sprayed areas should 1 receive consideration. 0 (5) Type of bottom for planting stations. It would seem advisable, where feasible, to plant fish in natural habitats. For young salmon this means in shallow rapids with gravel to rubble bottom rather than in deep pools. 17

(6) Eels a danger. Eels feed voraciously on newly-planted underyearlings. Therefore, planting sites should if possible be located away from areas of eel concentrations. Areas of large boulders and rubble in moderately swift water and pools are places where eels tend to be abundant. Eels are abundant in Bay of Fundy salmon B streams, but apparently not nearly so plentiful in Gulf of St. Lawrence salmon streams. No generally feasible means of controlling eels has so far been devised.

(7) Return of adults. The value of hatchery-reared stock, derived from other streams, for providing adult salmon to the stream planted, needs further measurement. Hatchery-reared stock do, however, contribute to sea fisheries, even far distant from the stream of planting. Acknowledgments Hatchery stock was supplied and delivered on the requested dates to designated points, by the Fish Culture Branch of the Department of Fisheries. Throughout much of the work I was able to call on Mr. H. C. White for advice and assistance in field problems. The unstinting efforts of the Pollett field crew, headed at first by H. W. Coates and later by P. R. Graves provided much of the data on which these studies are based. I wish to acknowledge my indebtedness to all who assisted in any way. Literature Cited

ELSON, P. F. 1957a. Number of Salmon Needed to Maintain Stocks. Canadian Fish Culturist, No. 21, pp. 000-000. ELSON, P. F. 1957b. The Importance of Size for the Change from Parr to Smolt in Atlantic Salmon. Canadian Fish Culturist, No. 21, pp. 000-000. GODFREY, H. 1956. Catches of Fish in New Brunswick Streams by Direct Current Electrofishing. Canadian Fish Culturist, No. 19, pp. 1-8. HOAR, WILLIAM S. 1939. The Thyroid Gland of the Atlantic Salmon. J. Morph. Vol. 65, No. 2, pp. 2S7•295. MILLER, RICHARD B. 1953. Comparative Survival of Wild and Hatchery-reared Cutthroat Trout in a Stream. Trans. Am. Fish. Soc. Vol. 83, pp. 120-130. WHITE, H. C. 1939. Bird control to Increase the Margaree River Salmon. Bull. Fish. Res. Bd. Canada, No. 58, 30 pp.

t 0

Number of Salmon Needed to Maintain Stocks by P. F. Elson

Fisheries Research Board of Canada, Biological Station, St. Andrews, N.B.

Most salmon streams obtain the greater part of their stock through natural spawning. . How many fish must enter a river in order to keep up the stock? Such information can contribute materially to wise use of our Atlantic salmon resources. The number needed will depend not only upon the size of the river, but also upon how many young fish the river can raise by reason of its suitability for salmon, the abundance of predators and so on. Useful information on this problem is accumulating from research on the Pollett River, a branch of the Petitcodiac system, N.B., and on certain tributaries of the Miramichi River. This account is in the nature of an interim report pending accumulation of more complete data.

(a) HOW MANY SPAWNERS ARE NEEDED? Experiments on planting hatchery underyearlings (Elson, 1957a) have shown how many young salmon one particular stream could support. The Pollett River could produce only five to six two-year smolts per 100 square yards of stream bottom. Such production could be obtained by planting underyearlings at the rate of about 35 per 100 square yards. This in turn should give about "10 large" parr, i.e. over 10 cm. or four inches total length measured from tip of snout to tip of tail (Elson, 1957b) which is the number required to assure the maximum production of smolts. Small salmon fingerlings, when first taken from a hatchery and liberated in a stream, do not live quite so successfully as native, wild salmon of the same size. But when they have lived in the stream for a year or more, they are probably equally valuable for producing smolts. So the question may be stated-How many adult salmon are needed to give large parr at a rate of about 10 per 100 square yards? Studies are now being made of the numbers AF underyearlings, parr and smolts produced by known numbers of adult salmon entering a 10-mile stretch of the Pollett River. The number of eggs brought into the stream is of more importance than the number of fish bringing them. Pollett salmon, most of which are grilse, weigh on the average, about 44 pounds. The number of eggs carried by female salmon tends to be in proportion to the weight of the fish. Pollett salmon carry 800 per pound of live weight. Hence it is useful to think in terms of the total weight of female salmon in the river, rather than their total numbers. In general, sex ratios and spawning habits of salmon are such that there tends to be a sufficient number of males for the females. The results which have accumulated from these Pollett observations to date are summarized in Table I. 19 20

Table I

Underyearling and pre-smolt parr produced from known quantities of adults entering a 10-mile section of the Pollett River (25 yd. wide) based on seining in the same 10 sample areas each year. (a) Five to 15 per cent of large parr in the Pollett River are two-year-olds. This, or the limits of accuracy for seining could cause the discrepancy in the survival of underyearlings to parr.

Total weight of Potential egg deposition Underyearlings in large parr in 2nd summer adult females C;800 eggs/lb. of female 1st summer Survival from Total No. per No. per Survival No. per Year Pounds under- number 100 ^ yd. 100 sq.y d. from egg s 100 sq' yd. yearlings eggs

1953 210 168,000 38.6 2.2 (a) 6% 2.3 (a) 100 17C 6 % 1954 390 312,000 71.7 5.9 8% 3.2 54% 4% 1955 310 248,000 57.0 4.0 7% (in'57) - -

Using the results shown in this table, we can calculate that to get the 10 large parr needed in the Pollett for maximum smolt production will require, at the average survival rate of five per cent, potential egg deposition at a rate of about 200 per 100 square yards. This means about 44 pounds of female salmon (say 10 female grilse or four larger hen salmon) per mile of stream 10 yards wide. This is a general rate for the stream as a whole, not just for nursery areas or spawning beds. In 1956 the average potential egg deposition in the same area was 252 eggs per 100 square yards. This last spawning will therefore provide a good test for the above hypothesis.

(b) DO MARITIME SALMON STREAMS RECEIVE ENOUGH STOCK? There is strong public opinion that the Maritime salmon resource could be greatly increased by providing more very young stock, whether by legislating for greater spawning escapement, or by providing for greater distribution from hatcheries. The Miramichi is frequently referred to as a stream which should receive such benefits. For several years now, measurements have been made of the adult salmon runs into two Miramichi tributaries, and of the resulting populations of young salmon. The Northwest Miramichi has about 70 miles of effective salmon rearing water, which averages about 25 yards wide; the Dungarvon has 50 miles, and a similar width. Examination of adult salmon runs has shown the average size of the fish and the proportions of males and females. Records of the Fish Culture Branch of the Canadian Department of Fisheries show that, in general, as for Pollett fish, about 800 eggs are carried for every pound of female salmon. From this information it is possible to estimate the number of salmon eggs brought into these two streams-the potential egg deposition. The average annual figures are, for the Northwest Miramichi nearly 72 million eggs and for the Dungarvon nearly 42 million. In Figure 1 potential egg deposition can be compared with resulting populations of young salmon as they advanced in age. All measurements are in terms of average numbers per 100 square yards of stream bottom. The values for young fish were determined by seining with electrofishing in several sample areas of each stream. 21

Tests have shown that the population figures for the sampling areas can be trusted to within about 20 per cent, ofter 10 per cent. Minor discrepancies in the figures given can be attributed to such error in counting the fish, but the general trends can be regarded as well established. For the Miramichi, nearly all small parr (10 cm. and under) are yearlings and over 90 per cent of the large parr (over 10 cm.), which will become smolts the following spring, are two-year-olds (Elson 1957b). For the diagram, parr have been allocated to years of egg deposition on the basis of size. Age analysis of samples, from studying scales, will doubtless result in minor corrections to the data. NORTHWEST DUNGARVON MIRAMICHI

300

Ul u 200 W 100

'50 '5S 'SI I C } •O o W s0 n RENOUS CAIN S

4 W a 20 10

OC W 'S2 'S7 'S5 'SS CO 1 I 1 I 7 I I z 40 I I I Cr I I I CL 30 W ^ + + 0

'S2 'S7 '53 I I I I 1 ^ 10

L 1 wu C=ml 1.'R_ '53 'Se '54 . 'S5 'S7 '57

Figure 1.-Abundance of young salmon of various stages found in Miramichi streams by systematic population studies. Vertical arrows ( i ) indicate survival between different stages of one year class. White columns (_q) indicate abundance under natural condition. Dotted (_) and black columns indicate abundance of groups affected by merganser control. Black columns (A) indicate abundance of groups subjected to DDT spraying as well as bird control.

Recall that in the Pollett River 10 large parr per 100 square yards were required to assure maximum smolt production. In the Northwest Miramichi similar numbers of large parr resulted from spawnings in 1949 and 1950 (Fig. 1). These two groups 22 received full benefit of merganser control; earlier groups did not and later groups were reduced as a result of spraying the drainage basin with DDT against an epidemic of spruce budworm (Kerswill and Elson, 1955). However, the two groups do provide evidence that the Northwest was then receiving a good supply of young stock. On the Dungarvon, full production of large parr was not realized in the absence of mer. ganser control. However, underyearlings were as abundant as on the Northwest- Under natural conditions, potential egg deposition at a rate of 300 per 100 square yards did not produce many more underyearlings (fry) in these two streams than did potential egg deposition at 200. (Compare Northwest and Dungarvon in 1952 and Northwest in 1950 and 1951, in Figure 1.) Such a condition carries the implication that a potential egg deposition of 200 eggs per 100 square yards did not limit the production of young salmon in these streams. The Renous and Cains Rivers carried just as good stocks of underyearlings. (Fig. 1). This evidence indicates that the Miramichi has recently been receiving all the young stock that it can rear to more advanced stages. How does the potential egg deposition in these streams compare with the calculated requirement of 200 eggs per 100 square yards made for the Pollett? The average for the Northwest has been about 235 and the Dungarvon 200; the weights of female salmon received by these Miramichi tributaries are respectively 52 pounds and 44 pounds per mile of stream 10 yards wide in comparison with the 44 pounds calculated to be required for the Pollett. There is some possibility that the Miramichi may not actually require as many eggs per 100 square yards as the Pollett in order to have the 20 underyearlings apparently necessary to give highest production of smolts. For one thing, eels have been found to prey on both salmon eggs and alevins still in the redds, as well as on older stages. But eels are only from one-tenth to one-twentieth as abundant in the Miramichi as in the Pollett. The average natural survival rate from eggs to fry found for the Miramichi (8 percent) was just as good as that found on the Pollett (7 per cent) with very light spawning. A very light spawning should give a maximum survival rate, since no excess over actual requirements is involved. In the Miramichi area, the lighter spawnings observed did give the best survival rates so some excess over actual requirements may well have occurred in the heavier spawnings. During the period of these observations the Miramichi has received some plantings of underyearlings salmon each year. The Northwest Miramichi, has been planted with underyearlings at an average rate of 25 per 100 square yards for the whole stream. All plantings were actually placed in the lower 25 miles, which are accessible for hatchery distribution. Young salmon were most abundant in the upper 40 miles. On the Dungarvon, plantings provided underyearlings at an overall rate of six per 100 square yards; but all were liberated in the lower five miles of the stream because above this it is impractical to reach the river with hatchery products. Again, young salmon were relatively much more abundant in the upper part of the river. Thus the largest parts of these streams were not available to hatchery stocks (Elson 1957a). Most of the young stock must have come from natural spawning. It does appear that several Miramichi streams have, in recent years, received sufficient young stock, mostly native, to produce many more smolts than they are now 23 doing. In fact, natural spawning has been sufficient to give maximum smolt production, if judged by standards developed on the Pollett River. It is hard to see how such streams could, under normal conditions, use to advantage more brood stock or hatchery products than they have recently been getting. With the information at present available the best figure for the number of adult salmon required to maintain stocks can be set at "between 40 and 50 pounds of adult females per mile of stream 10 yards wide". However, as shown in Figure 1, these young stocks are now seriously reduced as a result of spraying the adjacent forest with DDT. This chemical kills a high percentage of the young salmon. If spraying is to continue, natural spawning alone cannot provide a remedy, since the young fry are most susceptible of all to the spray. Care- fully planned use of hatchery stocks seems to be the one way of getting better production under these circumstances.

Acknowledgments Much of the data used has been collected through the untiring efforts of the Pollett field crew, consisting of P. R. Graves, H. P. Barchard, L. MacFarlane and A. G. Steeves, of Elgin, N.B. Others assisted this group in the Miramichi seining studies. Dr. C. J. Kerswill provided data on Miramichi salmon runs. To these and all others whose assistance facilitated the work I extend my thanks.

Literature Cited

ELSON, P. F. 1957a. Using Hatchery-reared Atlantic salmon to best advantage. Canadian Fish Culturist, Issue VI, PP. 00-00. 1957b. The importance of size for the change from parr to smolt in Atlantic salmon. Canadian Fish Culturist, Issue VI, pp. OOAO. KERSWILL, C. J., and P. F. ELSON. 1955. Preliminary observations on effects of 1954 DDT spraying on Miramichi salmon stocks. Fish. Res. Bd. Canada, Atlantic Prog. Rep. No. 62 pp. 17-2

The Role of Hatcheries in Assuring Maritime Stocks of Atlantic Salmon by P. F. Elson

Fisheries Research Board of Canada, Biological Section, St. Andrews, N.B.

The abundance of Canadian Atlantic salmon as measured by records of catches has varied a good deal from time to time, even in "the good old days." Man has attempted to reinforce the stocks of these fish by artificial propagation of the young. Canada was one of the foremost countries in bringing such culture to a high stage of development. Our system for rearing young salmon and distributing them in suitable streams is one of the largest in the world. Extensive distribution of hatchery-reared salmon in eastern Canada began about 1875. The early distributions were followed by large increases in the catches of salmon after the expected intervals, and in the proper districts (Wilmot, 1885). These increases were very welcome, following, as they did, right on the very lowest commercial salmon harvest on record, in 1881 (Fig. 1). Hatchery distributions continued to increase until they reached about 25,000,000 fingerlings yearly in the 1930's. Since then they have been reduced to around 10,000,000 to 15,000,000.

RELATION OF HATCHERY DISTRIBUTIONS TO CATCHES:

Have these widespread plantings had the notable effect on salmon harvests that was predicted in the 1885 report? If so, years of heavy planting should have been followed in due course by particularly good catches of salmon, and smaller plantings by smaller catches. For the Maritime Region (Cape Gaspé to the St. Croix River) most commercially caught salmon spend three years in rivers and two years at sea. The common interval from planting to harvest is therefore five years. The sizes of hatchery distributions and corresponding catches in this region have been arranged diagrammatically in Figure 2. In the middle row vertical bars representing hatchery distributions between 1907 and 1947 have been arranged in descending order of magnitude. In the lower row bars representing commercial catches made five years later are arranged directly under their corresponding distribution bars. There is no clear resemblance between the patterns formed by the two different rows of bars. The conclusion seems to be that the contributions of hatchery plantings to the general salmon catches have not been large enough to show through the natural variations in these catches.

Hatchery distributions might be compared with commercial catches made one year before the distributions. Such catches represent, more or less, the availability of parent stock to the hatcheries. These catches are shown in the upper row directly

25 26

ATLANTIC SALMON

LANDINGS

MIIhIIMIIUM ^1930 1940 1950 to"1880 1890 1900 1910

500 B 400 PARENTS

Z 3001i

LL 11111 1111111111111111111 O ^ Z e00 OFFSPRING ^ 400- D 0 = 300 f 200

100

Figure 1.-Commercial landings of Atlantic salmon from the Gulf Area of the Maritime Region for 85 years. A. Annual landings in chronological order. B. Same catches plotted as numbers of 10-lb. salmon. Upper panel: numbers of parents in the first 79 years arranged in order of magnitude. Lower panel: numbers of offspring caught six years later, drawn directly below corresponding parent catch. There is no evidcnce of any relation between the size of catches from parents and the size of catches from their offspring. (Reproduced from Elson, 1955) above their corresponding hatchery distribution. There is, in fact, a tendency for hatchery distributions to be larger in the years following abundant harvests of salmon. The relationship is obscured somewhat by a few particularly good catches; but the upper row of bars also tends to slope downward from left to right. It is quite logical that the hatcheries should be able to collect more eggs in years when mature salmon can be caught more readily. This is apparently what happened. Both of the relations discussed above have been tested mathematically and the conclusions presented found to be valid. As can be seen in Figure 1, there is no clear-cut relationship between the amount of parent salmon caught in one year and the amount of their offspring caught six years later (Elson, 1955). This can be interpreted as meaning that having a great many more salmon spawn in one year does not necessarily lead to an abundant catch of their offspring. Apparently naturally produced increases of young often do not show up in later catches. It should not be surprising, therefore, that arti5cally produced, smaller increases in young are not apparent either. 27

Thus the evaluation made in the 1880's of possible hatchery contributions has not been substantiated by later experience.

WHAT OUR HATCHERIES CAN CONTRIBUTE.

The fact that hatchery contributions cannot be identified by their effect on catches does not mean that they are not of value. In the last century practically nothing was known about the usual fate of planted fish. In the past quarter of a century considerable information on this subject has been collected. In those early years people

^ ^ COMERCIAL LANDINGS z D o PARENTS a

z 2 J I

i ^ âillkàâd 11111 Il i u z J cr HATCHERY DISTRIBUTIONS w 0 PROGENY

201 U. O ^ Z 10^ O_ J _J ^

h t]z COMMERCIAL LANDINGS =1 0 PROGENY

, J

a 1 Figure 2.-HatcheryuJd distributions of young Atlantic salmon arranged in order of magnitude (middle row) for comparison with commercial landings one year earlier (above) and five years later (below). There is a similarity of trend between earlier catches and distributions: this probably reflects availability of parent stock to hatcheries. There is no simi• lanty between the trends for distribution and catches flic years later when the planted fish should have made their biggest contribution to fishing: this indicates that hatchery contributions were not large enough to outweigh natural causes of fluctuation. 28

were pleased or worried over the status of salmon stocks according as recent catches had been good or bad (Perley, 1852) but there is no record of any attempt to estimate the total extent of Maritime salmon stocks. Now, as a result of researches made in the last 25 years, at least a rough estimate can be made. Thus it is now possible to place a better measurement upon the contribution which our hatcheries can make. For this the first question to be asked is: How many mature salmon could be produced from 15,000,000 young salmon liberated each year by the hatcheries? It has been calculated on experimental evidence (Elson, 1957a) that with careful planning and management up to 15 per cent of hatchery-distributed underyearling salmon can be expected to survive to the smolt stage. Under adverse conditions the survival rate is liable to be only a fraction of this amount, perhaps around three per cent. Sea- ward-migrating smolts have been counted and marked in a number of instances, and their numbers compared to the numbers of returning adults. From these and similar studies it is found that, on the average, about eight per cent of the smolts (Table I)

Table I

Estimated survival rate of salmon marked as descending smolts based on reported returns to fisheries and counting fences in rivers; except Hayes (1953), which is based on estimates of total smolt production and total salmon associated with the river.

Place Author Survival rate

% LaHave, N.S ...... Hayes, 1953 ...... 17 Margaree, N.S ...... Huntsman, 1941 ...... over 2 Miramichi, N.B ...... Kerswill, 1957 ...... 5 Moser River (Mill Brook), N.S ...... White, 1940 ...... over 10 Moser River (main stream), N.S ...... White, 1940 ...... over 5 Moser River (Mill Brook), N.S ...... White, 1941 ...... 10 Moser River (main stream), N.S ...... White, 1941 ...... 9 Moser River (main stream), N.S ...... White, 1943 ...... over 9 Apple River, N.S ...... White and Huntsman, 1938...... over 3

Mean survival rate and its standard error ...... 8 ± 1.5% survive the rigours of life in the sea and return to fisheries and rivers as mature salmon. Considering these survival rates, the yearly contribution of hatchery fingerlings could conceivably add about 180,000 adult salmon to the stock available for fisheries and spawning. This amount is roughly equivalent, in numbers, to the total take by commercial fisheries in the Maritime region.

MAGNITUDE OF SALMON STOCKS IN THE MARITIME REGION

1. Direct estimate from fisheries. In order to evaluate the hatchery contribution we should develop at least a general idea of the size of the total stock. Our impressions of abundance are usually obtained from the numbers of salmon caught in fisheries. But how is catch related to actual abundance? There have been a number ofstudies in which salmon caught in nets, at sea or along the coast, have been tagged and immediately released again. Subse- quent recapture of some of these fish leads to our best available estimate of the total 29

salmon stock. Table 11 gives the results from various areas. Applying the average value for recaptures (27 per cent) it appears that the combined commercial and sport fisheries of the Maritime region remove a little over one quarter of the salmon stocks available to the area. In recent years the commercial catch has amounted to around

Table II

Atlantic salmon tagged and liberated from commercial fisheries and recaptured in both commercial and sport fisheries.

Total Place of tagging Number tagged Author recaptures

Petit Gaspé, P.Q ...... 100 Sept Isles, P.Q ...... 80 Rivière Nabissippi, P.Q ...... 43 St. Augustin, P.Q ...... 63 Belding and Prefontaine, St.Paul P.Q ...... 150 1938 Port aux Basques, N9d ...... 599 Miramichi, N.B...... 411 Margaree, N.S ...... 758 Hayes, 1948 Margaree, N.S ...... 416 Hayes, 1949 LaHave, N.S ...... 192 Hayes 1953 Margaree, N.S ...... 100 Margaree, N.S ...... 161 ^Huntsman, 1939 Margaree, N.S ...... 267 Saint John, N.B ...... 102 Lorneville, N.B ...... 100 Huntsman, unpub. Dipper Harbour, N.B ...... 102

Total number tagged ...... 3,644

Mean rate of recapture and its standard error 27 ± 4.5^^

150,000 salmon per season. The sport fishery in this region has been estimated to take 55,000 fish yearly, of which about one half are grilse (Kerswill, 1957). The grilse must be considered because they represent young stock which has survived to maturity just as much as larger salmon do. The total catch thus amounts to about 200,000 fish. If, as indicated in Table 11, this is about one quarter of the available stock, there must be in the vicinity of 1,000,000 salmon available to the Maritime region. The rates of recapture listed indicate that it is unlikely that the total stock falls as low as 500,000 or exceeds 1,500,000 salmon in any one year.

2. Stock size at various stages In order to appraise the usefulness of hatcheries, it is helpful to have estimates of the salmon stocks at various stages from eggs to adults. Estimates of this kind can be obtained from our data by using a slightly different chain of facts. The total catch by fisheries, say 2,000,000 pounds, is still used as the starting point. Also the proportion (27 per cent) that this forms of the total stocks as indicated by tagging results must be considered. Now let us assume that more fish are lost to seals, sharks, poachers and other causes of deaths in the sea and rivers, to the extent that only half of the total mature stock in the sea is left in the streams for spawning. As a matter of fact, Hayes' data for the Margaree (1949) shows not over 58 per cent of the counted, 30 available fish for the system being thus available for spawning. (He counted fish removed by the commercial fishery, fish entering the river and fish removed by anglers.) Using the 50 per cent figure, we should have close to 4,000,000 pounds of salmon available for spawning in Maritime region rivers. In general, half of these would be females. Female salmon carry about 800 eggs per pound of body weight. The survival rate from eggs, brought into one particular river, to fingerlings the following summer has been established as about 5 per cent when there is no excessive spawning (Elson, 1957b). Using these figures, Maritime salmon streams, as a whole, are now receiving natural stocking at the rate of about 80,000,000 salmon fingerlings per year. Information now accumulating indicates that under favourable conditions, which include protection from excessive predation by mergansers, one-quarter or more of these fingerlings can survive to the smolt stage. These smolts, granting an 8 per cent survival rate over their sea life (Table I) would contribute a little over 1,500,000 salmon to the stock available for fisheries, spawning, etc. This second estimate is of the same general order as that obtained by a slightly different chain of thought, above. The tagging studies referred to in Table II were made in areas of relatively important fisheries; and of course only salmon which had already entered the fisheries could be tagged. There are considerable numbers of salmon associated with small streams where fisheries are less intense. There are also many salmon which apparently do not reach fishery areas during open seasons; for example, the extensive autumn runs of the Miramichi system. To the extent that the tagging studies take no account of such fish the estimate of total stock given above is too low. However, plans for reasonable exploitation and conservation of the stocks based on low estimates are quite unlikely to damage our salmon resource.

Discussion The potential value of the annual hatchery contribution might be evaluated on the basis of (1) 15,000,000 hatchery fingerlings added to the native stock of around 80,000,000 native fingerlings; or (2) 180,000 mature salmon added to the sea stock of about 1,000,000 fish; or (3) 50,000 salmon added to the 200,000 taken by fisheries. However assessed, the potential contribution seems to be of the general order of one- fifth to one-quarter of the present stock. An addition of 500,000 pounds (commercially caught salmon average about 10 pounds weight) to Maritime region catches is a valuable contribution. But, spread over the entire region, it could scarcely be identified in catch statistics of a fishery which has fluctuated from under 1,000,000 to over 5,000,000 pounds per year during the last 80 years. This explains, at least in part, why the value of hatchery contributions does not show in such an analysis as is illustrated in Figure 2. However, an additional catch of 500,000 pounds divided among a few smaller fisheries could make a very evident contribution to these. In the same way, an increase in general spawning stock of 75,000 to 100,000 salmon would scarcely be noticed by its effect. But this is enough salmon to completely stock a number of small rivers, or even one fairly large system. As an example, an 31

annual run of 3,000 to 4,000 grilse and salmon combined appears to provide good angling and sufficient spawning stock in the 70-mile-long Northwest Miramichi River. Thus Canada's hatchery system for Atlantic salmon can add a valuable amount to Atlantic salmon stocks. But even more important, it is a resource which can ensure against extinction of stocks if and when this should threaten. The professional fish culturist, alone, is likely to have ready access to the background of facts which are necessary in order to make a sound decision about where and how hatchery stock can be used to best advantage. He must first get some clue as to why fish are scarce. Could it be simply that weather and water conditions were not right for getting the salmon into fishery areas? Or are young salmon really scarce in the nursery streams? If so, are they scarce, from the very youngest stages up, or only the older stages? What are the conditions that face the different stages of young in any particular stream? To be most effective, hatcheries must be able to add fish to replenish the stage at which native fish are scarce in the stream. Hatcheries are not designed to remedy shortages by supplying excessive stock in advance of depletion and regardless of the cause. These are the sort of problems that must be considered in order to get the most out of hatcheries for Atlantic salmon. Whenever well- considered answers can govern their procedures, hatcheries provide, over the years, a means of bolstering and restoring damaged fisheries; of revitalizing fisheries which have nearly or entirely disappeared; and even of creating new stocks for new fisheries. But if uninformed demands or other circumstances govern their procedures, our hatcheries can, at best, hope to add only about one quarter to the value of the general salmon harvest. No contribution of such magnitude could be positively identified in catch statistics. Literature Cited

BELDING, D. L. and G. PREFONTAINE. 1938. Etudes sur le saumon de l'Atlantique. (Salmo salar L.).-1. Organisation et résultats généraux des recherches dans le golfe SainbLaurent en 1937. Con- tributions de l'Institut de Zoologie de l'Université de Montréal, No. 2, pp. 1-50. ELSON, P. F. 1955. Have Atlantic Salmon Been Overfished? Fish. Res. Bd. Canada, Atlantic Prog. Repts. No. 63, pp. 13-15.

1957a. Using Hatchery-Reared Atlantic Salmon to Best Advantage. Canadian Fish Culturist, No. VI, pp. 00-00.

1957b. Number of Salmon Needed to Maintain Stocks. Canadian Fish Culturist, No. VI, pp 000-000.

HAYES, F. R. 1948. Report of the Director of Fisheries. App. 1, Pt. II Margaree River. Ann. Rept. Dept. Trade and Industry, Nova Scotia, pp. 115-125.

1949. Report of the Director of Fisheries. App. 1, Pt. II, Margaree River. Ann. Rept. Dept. Trade and Industry, Nova Scotia, pp. 119-130. 1953. Artificial Freshets and Other Factors Controlling the Ascent and Population of Atlantic Salmon in the LaHave River, Nova Scotia. Bull. Fish. Res. Bd. Canada No. 99, 47 pp. HUNTSMAN, A. G. 1939. Salmon for Angling in the Margaree River. Bull. Fish. Res. Bd. Canada, No. 57 75 pp. 1941. Cyclical Abundance and Birds versus Salmon. J. Fish. Res. Bd. Canada, 5 (3) pp. 227-235. 1954. Management of Saint John Salmon. (Typewritten MS.) KERSWILL, C. J. 1957. Regulation of the Atlantic Salmon Fisheries. (Typewritten MS.) PERLEY, M. H. 1852. The Sea and River Fisheries of New Brunswick. Report of Her Majesty's Emigration Officer at Saint John, New Brunswick, Fredericton, 294 pp. 32

WHITE, H. C. 1940. The Moser Salmon Run. Fish. Res. Bd. Canada, Ann. Rept. Atlantic salmon and trout investigations, App. 17, pp. 30-31 (Mimeo.). 1941. The Salmon Run. Fish. Res. Bd. Canada, Ann. Rept. Atlantic Salmon and Trout Investi- gations, App. 25, pp. 31-32 (Mimeo.). 1943. The Return from Marked Smolts at Moser River. Fish Res. Bd. Canada, Ann. Rept. Atlantic Salmon and Trout Investigations, App. 23, p. 35 (Mimeo.). WHITE, H. C. and A. G. HUNTSMAN. 1938. Is Local Behaviour in Salmon Heritable? J. Fish. Res. Bd. Canada, 4 (1), pp. 1-18. WILMOT, SAM. 1885. Report on Fish-Breeding in the Dominion of Canada, 1884. Canada Dept of Fisheries. First Ann. Rept. Supplement No. 2, pp. 1-71. EDMOND CLOUTIER, C.M.G., O.A., D.S.P. QUEEN'S PRINTER AND CONTROLLER OF STATIONERY OTTAWA, 1958 ISSUE TWENTY-TWO MAY - - 1958

THE CANADIAN FISH CULTURIST

Pubhshed at Ottawa by The Department of Fisheries of Canada

LIBRARY FISHERIES AND OCEANg BIBLIOTHEQUE 'PÈCJIES ET OCÉANS

EDMOND CLOUTIER, C.M.G., O.A.. D.S.P. QUEEV'8 PRINTER AND CONTROLLER OF STATIONERY OTTAWA, 1958 CONTENTS

Page Some Principles Involved in Regulation of Fisheries by Quota- W. E. RICKER ...... 1

Regulation of the Atlantic Salmon Fisheries-C. J. KERSWILL. 7

Regulation of the Lobster Fishery-D. G. WILDER ...... 13

Some Sociological Effects of Quota Control of Fisheries- J. L. HART ...... 17 Some Economic Aspects of Control by Quota- W. C. MACKENZIE ...... 21

The papers in this issue were presented at a meeting of the Com- mittee on Biological Investigations of the Fisheries Research Board of Canada held at Ottawa in January, 1957, under the chairmanship of Dr. J. R. Dymond, of Toronto.

The Canadian Fish Culturist is published under the authority of the Minister by the Department of Fisheries of Canada as a means of providing a forum for free expression of opinion on Canadian fish culture. In the areas of fact and opinion alike, the responsibility for statements made in articles or letters rests entirely with the writers. Publication of any particular material does not necessarily imply that the Department shares the views expressed. In issuing The Canadian Fish Culturist the Department of Fisheries is acting only as an instrument for assisting in the circulation of information and opinion among people in the fish culture field. Those who may wish to discuss articles which have been publiehed in The Canadian Fish Culturist are encouraged to do so and space will be made available.

Correspondence should be addressed to the DIRECTOR, INFORMATION AND EDUCA- TIONAL SERVICE, DEPARTMENT OF FISHERIES, OTTAWA, CANADA.

Published under Authority of HON. J. ANGUS MACLEAN, M.P., Minister of Fisheries Some Principles Involved in Regulation of Fisheries by Quota by W. E. Ricker

Fisheries Research Board of Canada, Biological Station, Nanaimo, B.C. All kinds of fishery regulations are designed to stop people from catching fish, even when they promise larger catches later on as a reward. Regulations may operate in one of a number of ways: 1. They may reduce the catch of a species at all sizes, or only at some particular range of sizes; 2. They may prohibit fishing in certain places; 3. They may prohibit fishing at certain times of year, or certain times of the day; 4. They may prohibit the use of certain kinds of gear; 5. They may prohibit taking fish in excess of a certain number (or weight) by any one individual, or boat; 6. They may prohibit taking fish in excess of a given total number or weight, each year, from the stock in question. The motives which lead to these regulations are all, I believe, economic in the broad sense. However they are classifiable in two groups: those which aim at direct economic ends, and those which aim at increasing the general level of potential catch. Among the more direct economic aims may be, for example: 1. a desire to share the catch among as large a number of fishermen as possible (e.g., bag limits in sport fisheries); 2. a desire to restrict the number of individual fishermen or boats so that each can have a desirable net return from his activities (for example, leasing of oyster beds, and restricted licencing of nets or traps); 3. a desire to favour certain types of fishing, or certain groups of fishermen (prohibition of trawling by large trawlers within 12 miles of the coast, in the Maritime Provinces); 4. a desire to spread out landings over the year to promote orderly marketing of fresh fish (separate "summer" and "winter" quotas for prairie lakes); 5. a desire to restrict the duration of fishing so as to avoid stormy weather (prohibition of winter halibut fishing in the north Pacific) or so as to work only when the fish are concentrated and easy to catch; 6. a desire to obtain a more valuable product by increasing the average size of fish caught (special size restrictions on lobsters in certain areas), or by catching fish only when they are of prime quality (for example, salmon or herring before their fat is converted to sexual products). The second type of regulation, often called biological, has to do with obtaining maximum sustained yield from the stock, or some similar objective. Regulations of this type are distinguished by the fact that they try to promote the economic objective of increased yield, or a larger catch per fisherman, by way of a manipulation of the size of the stock or its age composition. In their economic aspects, these so-called "biological" measures tend to benefit all fishermen indiscriminately, rather than the interests of particular groups. They are often described by names that have a certain emotional content, like "protection" or "conservation"' or "prevention of depletion". 11 am not, of course, deprecating conservation. Using the word in its strictest sense, there is a place and a need for conservation of various fishes and of other animals and plants from the forays of man, almost everywhere on the globe.

1 53562-5--1 2

The actual kinds of biological regulation vary somewhat. With salmon, which in Canada are mostly caught only toward the end of their growth, effort is devoted mainly to getting the best number of spawners on the redds-not too few and not too many-and getting them there at the proper time. With species like cod or flatfishes, where the individual fish can be available over several years of its life, the problem of maximizing yield has two aspects. For any given number of fish entering a fishery, and given amount of fishing gear, a best minimum size of fish caught can be computed from existing rates of growth and natural mortality, which will permit maximum catch from that group of fish. Existing restrictions on net size in the North Sea and parts of the northwest Atlantic are based mainly on computations of this type. However, the amount of recruitment, also, can vary with stock size or average fish size, as can rate of growth and possibly natural mortality rate; and these effects may modify or might even reverse the expected catch trend computed from existing conditions. We need to consider not only best use of available recruits, but also maintaining and increasing the production of recruits; and compromise may be necessary among conditions which favour each goal. Before leaving these generalities, we might notice that a given regulation may sometimes serve both direct economic purposes and also the less direct economic goals which have a biological background. This is fair enough, as long as there is agreement that both ends are desirable. What is not so desirable, but quite common nonetheless, is for the biological aspects of a regulation-the "general good"-to be made a front for promoting the interests of particular groups of fishermen. I recall how the salmon traps of Puget Sound were abolished in 1935 by a public referendum or "Initiative" in the State of Washington, which had the combined support of the salmon seiners and the sport fishermen. Among the arguments put forth was the contention that the traps were a particularly destructive and pernicious type of gear, that they "didn't give the fish a chance", though actually they were taking far less salmon than the seines did. In the same vein, sea fishermen will maintain that it is shameful for salmon to be caught after they get into a river, and river anglers consider it positively sinful to take salmon from spawning beds-though the salmon are just as dead when caught in one place as in another. It would be foolish to conclude from this that all kinds of fishing are equally desirable, from the point of view of economics and public policy: however, each method should be rated on its own relative contribution to income and to satisfaction, rather than on the alleged destructiveness of competing methods.

Characteristics of Catch Quotas

So much for regulation in general. One method of regulation is to establish a catch quota, and at this meeting we are interested in quotas particularly. On the non-biological side, daily quotas for individual sport fishermen are a very popular device for sharing the wealth, though quite often they are ineffective because they are made so large that few fishermen ever reach them. Individual boat quotas are also sometimes applied in commercial fisheries, usually by mutual agreement rather than by law, in order to share markets or processing facilities in times of temporary glut. 3

However, the term quota usually refers to a limit set on the total catch taken from a stock during a season. What are the peculiar advantages and disadvantages of such quotas? Now it happens that one of the most spirited eulogies of quota regulation has been made by Professor F. I. Baranov (1947). He refers to quotas as the "American" system of fishery management, in contrast to the "European" system of closed seasons, closed areas, gear restrictions, etc. Furthermore, his knowledge of and affection for quotas is based on accounts of their operation in two Canadian fisheries- the whitefish fisheries of some of our prairie lakes, and the eastern Pacific halibut fishery which is shared by Canada and the United States. I can do no better for quotas than to quote some of his remarks. "All restrictions on fishing," says Baranov, "are useful only insofar as they lead, by one means or another, to a decrease in the catch. In that event it is simpler, more logical and more honest to limit the size of the catch directly, without stooping to petty interference with fishing techniques." Further, "once a quota is established, all other restrictions on the fishery merely increase costs for the operators, without doing the fish stocks any good." "The work of regulatory authorities is simplified by quotas) ... and it is possible to reduce the enforcement apparatus. The work of the fishermen is also simplified. They can use the greatest fishing power of their nets, and they can devote their means and energy not to multiplying fishing apparatus in pursuit of an impractical level of catch, but in a restriction and rationalization of this apparatus for attaining the highest economic return from the established quota." Contrasting quotas with other methods, Baranov notes that "properly speaking, the majority of traditional restrictions on fishing are also directed toward decreasing the intensity of fishing." "But," he says, "they do not achieve that goal. Prohi- bition of one kind of fishing results in the fishermen turning to a different, still-legal, one, so that the amount of fishing is not reduced. Establishing a limit for size of nets ... is followed by an increase in their number. Closure of one region to fishing is compensated by increased activity in another. Introduction of a closed season on fishing means that fishing becomes greater during the remainder of the year." There is obviously much truth in these various points, and it would be easy to document them by illustrations from many fisheries. Yet anyone who is close to fishery administration knows that regulation by quota is not always and everywhere as simple and as beneficial as Baranov's words make it sound. An important technical problem is the need for a highly developed statistical system which provides day-by-day information on catch taken, especially as the quota limit is approached. Such a system can be developed, and it has in fact been developed for the British Columbia herring fishery and salmon fisheries, for example; but it is expensive. If fishing success is more predictable, and if the number of boats that will fish can be estimated fairly accurately, then up-to-the-minute statistics are not essential; an estimate of the probable catch per day can be made in advance with sufficient accuracy, and the fishing season can be made of a length to conform to the desired quota. In some fisheries neither of the above methods of approximating a quota may be available, or there may be other conditions which make an overall quota dangerous or impractical. For example, a quota covering a broad area may lead to overexploitation 4 of nearby grounds and underexploitation of distant ones: since the fishermen, each competing for a share of the quota, will make their cruising time as short as possible. This can be alleviated by setting individual quotas for individual subareas; but when this is done there immediately arise those problems of supervision and enforcement which quotas are supposed to obviate. Still other special conditions may make the use of overall quotas unrealistic, particularly where successive populations of a migratory fish proceed through a fishing area. Furthermore, when you get down to cases, the difference between a quota system and other methods of regulation is by no means clear-cut. Even in Baranov's example of the Pacific halibut, we find that the actual regulations specify a certain fishing season or seasons, not a certain total catch for the year. The length of the season is of course estimated to conform to the desired catch; but, ideally at least, any regulation of season or time of fishing is made with a view to adjusting the catch to the most desirable level. Thus the difference between the "traditional" methods of regulation, and the so-called "American" method of regulating by quotas, becomes tenuous and indistinct. However l do not suggest that the distinction disappears altogether. One thing which a quota brings into sharp focus is the question of the productivity of the stock. If a quota is set at 50 million pounds, for example, the first question a fisherman asks is, why is it 50? Why not 60? or 40? For that matter, how do you know that there will actually be 50 million pounds of fish on hand? Such questions are pertinent with any kind of regulation. But when a numerical quota has to be announced, they have a special urgency; there is no avoiding them by vague reference to a "need for conservation" or "protection of spawners". I think we can fairly claim as an advantage for quotas, over other types of regulation, that they more urgently demand information on stock size and productive potential, and that they stimulate efforts to determine these2.

British Columbia herring quotas I have been asked to indicate the status of the quotas on herring catches in British Columbia, in relation to the general question of regulation by quota. A good general summary of this matter is given by Taylor (1955). Catch quotas were introduced to a part of the British Columbia herring fishery first during the 1936-37 season. Prior to this time important fishing areas east and west of Vancouver Island had been closed to fishing for herring to be used for reduction purposes-the feeling being that there was greater economic return in using the herring for human food. However the demand for food herring had decreased, and the Province consented to wholesale reduction provided a quota was set to avoid possible excessive exploitation. The quotas set were below the known maximum catches but greater than the minimum catches for the areas concerned. The supply of British Columbia herring varies from year to year, but usually only moderately: in this respect they are intermediate between halibut and salmon. Nowa- days only two or three age-groups occur in the catches to any important degree. Success

2Note especially that knowing the size of the stock is not enough. Too much preoccupation with size of stock can !ead to the conscious or unconscious assumption that a large stock is an end in itself. Both theo- retically and from experience we know that this is not so. It is possible for a stock to be too great to provide maximum yield, as well as too small. 5

of reproduction varies considerably from year to year-more in some stocks than in others, but nowhere nearly as much as in, for example, the Norwegian herring stocks. If there are two poor year-classes in succession it means poor fishing two to three years later, while successive good reproductions set the stage for an abundant catch. In addition, in certain areas the stock is more vulnerable to capture in some years than in others. Mainly because of this last, predictions of each year's supply are not yet reliable enough to be used to set a special quota in each area for each year in advance; instead there is nominally the same quota for all years. In practice, when herring are relatively scarce the quota is not taken. When they are abundant, nearly always a quota extension is granted. The result is that the quota sets a limit to the catch only about once in three years, and even when it does, usually only. a few days remain in the season and not much more herring could be taken anyway. However the herring stocks and fishery are far from being unregulated. There exists a closed season during spawning time, which protects the fish at a time when they are massed into very dense schools. Appropriate gear could make very large catches at such times, and could increase the total seasonal rate of exploitation of the stock from the present 45-55 per cent up to possibly 80 per cent, 90 per cent, or even higher. In summary, the British Columbia herring fishery is producing an excellent annual yield and gives every appearance of being in a healthy state. But it is not a good example of the effects of regulation by quota, since the quotas to-day do not often or very greatly restrict fishing: restriction is accomplished by the closed season. How- ever, the quotas may have potential value as a standby, in case pre-spawning fishing were ever to become substantially more effective than it is to-day.

Effects on fishermen and the industry

I do not intend to consider the economic and sociological effects of quota regulation in any detail. However it seems necessary to comment on one of the statements quoted earlier: that the use of quotas makes it possible for the fishermen to restrict and rationalize their fishing, by eliminating gear in excess of what is needed to take the quota. This of course is possible if quotas are set for individual fishermen or boats, and this system has been used in the whitefish fishery of, for example, Lake Winnipeg. A disadvantage, possibly, of this practice is that it may tend to discourage individual initiative among the fishermen (since the more skilful or enterprising ones can take no more than others) and it relaxes normal incentives to improve fishing equipment and to exploit fully the whole area from which a catch can be taken. On the other hand, if only an over-all quota is assigned and there is no limit on total gear in use, each fisherman is exposed to the full rigours of competition-something he may not altogether care for. The general economic situation is then the same as for other common-resource operations, whose advantages and disadvantages will be evident from the remarks of the speakers who follow. I

References

BARANOV, F. I. 1947. The American system of fishery management.1 Rybnoe Khoziaistvo for 1946, No. 12, pp. 31-34. TAYLOR, F. H. C. 1955. The Pacific herring (Clupea harengus) along the Pacific coast of Canada. Intern. North Pacific Fish. Comm., Bull. No. 1, pp. 107,128. Regulation of the Atlantic Salmon Fisheries by C. J. Kerswill

Fisheries Research Board of Canada, BiologicaZ Station, St. And7ews, N.B.

Since 1949 the research and management programme for Atlantic salmon in eastern Canada has been under review by a Federal-Provincial Co-ordinating Committee. The main objective is to make more salmon available for use by both the commercial and sport fishermen. The various projects and recent results of the joint programme have been described in articles published in the April, 1955, and April, 1956, issues of the Department of Fisheries' monthly publication "Trade News".

Economic Value of the Salmon Fisheries Wide fluctuations in total annual commercial landings of Atlantic salmon in eastern Canada have occurred since 1870 when the first statistics were published. Peak levels of production usually followed periods of below-average yields at intervals of about ten years. The most recent peak in commercial production occurred in 1930, however, when over 13 million pounds were landed, with a value of over $4 million. This was followed by a steady decline to a level of 4 million pounds in 1945, no indication of significant improvement by 1949, and a still lower level of 3 million pounds with a landed value of about $1 million by 1955. From one-half to two-thirds of these total landings have come from Newfoundland, including Labrador. The landings in all parts of the Canadian Atlantic coast have shown a similar trend, and it is interesting that there is a close correspondence between fluctuations in commercial landings here and in the eastern part of the North Atlantic around the British Isles. There has been widespread concern over this last prolonged depression in commercial landings with its widespread effect on the netsmen, and anxiety about the future of the valuable sport fishery in the rivers. Actually there is no evidence of a general decline in angling catches since 1930, and even more salmon have likely been caught through an increased total rod effort in recent years. Since 1949 improved statistics have been obtained on angling catches, and it is known that at least 75,000 salmon are now taken annually in all the rivers combined. Of these about 45,000 are angled in the rivers of the Gulf of St. Lawrence, 5,000 in rivers around the outer coast of Nova Scotia, 5,000 in Bay of Fundy rivers, 5,000 in Quebec rivers on the north shore of the St. Lawrence, and 15,000 in Newfoundland rivers. Unfortunately it is very difficult to assign a reliable economic value to the sport fishery, where the value of recreation alone accounts for such a large proportion of the total. Judging by the value of comparable sport fisheries in the United States recently estimated by a careful census, our salmon angling is probably worth several million dollars. There is no doubt that the availability of Atlantic salmon for angling in Canadian rivers is a very valuable national asset. This, combined with the commercial fishery, which can be 7 53562-5-2 8

expected to be worth about $3 million in normal years, justifies considerable effort to see that the best possible management techniques are employed for this species.

Salmon Life History in Relation to the Fisheries in a`I'ypical River Area The accompanying figure shows the life history of Atlantic salmon and its relationship to the commercial and sport fisheries in a typical large New Brunswick river system. Over the whole Atlantic coast of Canada, the duration of the fresh water stage varies more than is indicated for this one area.

Spawning A convenient starting point is the spawning of adults in October and November. Experiments on the Pollett River, N.B., are now indicating that a rather modest number of parents is needed for optimum smolt production If predatory birds, chiefly mergansers, are controlled it is estimated that about 45 pounds of adult females are needed per average mile of stream 10 yards wide, to give the optimum number of about 200 eggs per 100 square yards of bottom. This applies only to streams which commonly have two-year smolts. For rivers like the Miramichi with typically three- year smolts, about 250 eggs per 100 square yards seem to be needed. This allows for normal fishing effort after the adults have entered fresh water, so the numbers of eggs are higher than are actually required to be deposited in the spawning redds. Studies over the past five years indicate that the Miramichi River has been receiving about this level of spawners. An important point is that a much lower egg deposition can be allowed if mergansers are not controlled, because these predatory birds will drastically reduce the number of parr to give much lower smolt production than would otherwise occur. To allow for the above levels of spawning without predatory bird control would be expected to lead to a waste of adult salmon.

From eggs to smolts In the Pollett River experiments now in progress, the survival rate from eggs to fry with three "light" spawnings has averaged 6 per cent, while the survival rate from fry to large, pre-smolt, parr has been 80 per cent. Comparable data for "medium" and "heavy" spawnings are not yet available. It is hoped that the results of these studies combined with similar observations on other rivers will shortly give a relationship between smolt production, egg deposition, and numbers of spawning adults.

From smolts to adults There are inadequate data yet on the rate of survival from descending smolts to the adult stage when the salmon are taken by various fisheries or contribute to the spawning escapement. Some evidence that the survival rate lies between five and ten per cent has come from our smolt marking experiments and from other investigations. For the Miramichi River the estimated total annual smolt production in recent years has averaged about lZ million. With such a high mortality in the sea, only 75,000 to 100,000 of this output would survive for use by the fisheries or for spawning. ADULT SALMON PARR I TO 3 YEARS OLD 4 FINGERLINGS I SEA YEAR (UNDERYEARLINGE) (GRILSE) ^ FRY 2 SEA YEAR ETC.

%o 10

Capture by commercial gear For the whole Canadian coast the gear comprises surface drift-nets operated offshore, inshore fixed trap nets and a few floating gill-nets, while some salmon are taken incidentally in gear set for other species. Traditionally the nets have been set to take adult fish as they apparently move from the feeding areas which are often far away in the sea, towards the native streams to spawn. Reliable information on the capture of adult salmon of known river origin by such gear over a wide area of the coast is now coming from a marking programme involving smolts produced in the 4 Miramichi and Pollett Rivers, N.B., the Port Daniel River, Que., and the Little Codroy River, Nfld. It has been found, for example, that many salmon produced in Maritime streams are taken far away in commercial nets, particularly around the east coast of Newfoundland. Also they may wander into the estuaries of other maritime rivers and be caught there by commercial gear. Of course, many are taken also by the gear set in the estuary of their native river. There are fishery regulations covering the opening and closing dates of the seasons which are different for different areas of the coast, mesh regulations aimed at protecting small salmon (grilse), weekly close periods to allow fish to escape the nets completely for a couple of days each week, but no quota systems. Most of the catch consists of two-sea-year fish, except around Newfoundland where grilse can still be caught for market in commercial nets. Each year the run of fish into many estuaries is marked by two peaks, one early in the summer and one in the autumn. In some rivers like the Miramichi the late run has by far the most fish, which come in after the commercial season is over. Only early run fish are desired by most commercial fishermen, and at present there is no demand for a fall fishery. As the spawning season approaches, salmon are not considered to be in good condition for the market.

The sport fishery Upon reaching fresh water areas of the rivers the salmon become available to anglers, who can legally keep both grilse and large salmon. In some rivers, like the Miramichi, grilse comprise about one-half the total angling catch. Other rivers, like those of the east coast of Newfoundland, may provide only grilse angling, while in others large salmon provide the bulk of the angling catch as in some Quebec streams. In New Brunswick alone there is a legal fishery for kelts in the spring. There seems to be no objection to this on biological grounds because very few kelts (5 to 10 per cent) ever turn up again after the first spawning. About 5,000 kelts are taken each spring in the Miramichi River which is now the most popular place for such fishing. The fishery regulations for angling include opening and closing dates for the season which differ from one area to another depending on the usual time of arrival of the fish, daily and weekly bag limits, and restriction of method to the use of artificial flies when angling for salmon.

Discussion The primary aim of fishery regulations for Atlantic salmon is to assure an adequate spawning escapement. Obviously most of the existing regulations were put into effect many years ago when requirements for spawning were unknown. Reliable 11

information on this point is now being obtained for the experimental rivers where concentrated work is in progress. It is hoped that the requirements for many streams can be developed from data obtained on a few typical ones. A useful supplement to these studies is the counting of ascending fish through traps installed in many fishways, and annual surveys of spawning activities on a number of rivers now being carried out by the Conservation and Development Service of the Department of Fisheries. It seems reasonable to aim at providing, through regulation, for the amount of 4 spawn that will give the optimum number of young under average conditions for hatching and survival. An idea of the amount of variation to be expected from year to year in the smolt production of the same river area is provided by data of recent f experiments on the Pollett River. On a ten-mile experimental stretch, ample stocking was provided by planting 250,000 fingerlings in five successive years, with predatory birds controlled. The average smolt production from these five plantings was 20,000; the minimum was 14,000, the maximum was 25,000. This variation amounts to about plus or minus 25 per cent of the basic "capacity" of the stretch to produce smolts. A very important point concerning regulations to insure adequate spawning escape. ment, is the possibility of early-run or late-run adults tending to produce young salmon that will inherit the tendency to return to fresh water early or late, as adults. If this were the case and if attention were given only to the numbers of young salmon being produced per unit area of stream, the future supply of early-run adult salmon could be inadequate. There is some evidence that such a tendency to return early or late is not generally inherited in Atlantic salmon. Nevertheless, in 1954 the fishery regulations for both commercial and sport fishing were amended to allow more early-run salmon to reach the spawning grounds. It was reasoned that this could do no harm and that it was desirable to be on the safe side until more information could be obtained. This information will start to accumulate in 1957 as the first returns of adult salmon come from an experiment started in 1953. The experiment involves the planting of marked fingerlings of known parentage, in streams which have runs predominantly of the other type as regards time of entry of adults. The time of return of the marked salmon will be compared with the time of return of the native stock. Most of the fishery regulations for Atlantic salmon have likely been too restrictive in the past, owing to inadequate knowledge. An example of a situation where liberalization of regulations might be desirable is the closing date of the commercial fishing season in the Miramichi River area. Closing the trap-net season on August 31 prevents utilization of most of the total run of salmon here each year, because at least 85 per cent of all the large salmon enter the estuary after September 1. Since 1954, however, abnormal conditions affecting smolt production have been created in the whole Miramichi watershed through extensive spraying of the woodland with DDT against the spruce budworm. The spraying has had such serious effects on parr survival that it is not safe to reduce spawning stocks at this time. Several years ago when the present research and management programme was being planned it was necessary to assign priorities to a list of projects considered to be worthy of attention in the near future. Almost all the projects have been undertaken 12

except physiological and behaviour studies which would require special facilities and staff. Quite recently it has become apparent that particular attention should be given to the effects of hydro-electric developments, pollution as from mining operations or DDT spraying, deforestation, and other factors associated with industrialization, on the general behaviour and survival of various stages of salmon. New studies along these lines are now being planned. It is quite possible that the information so obtained will be of great importance in any attempts to manage the Atlantic salmon fisheries by regulation or other means a few years hence. Regulation of the Lobster Fishery

by D. G. Wilder

Fishertes Research Board of Canada, Bidogical Station, St. Andrews, N.B.

Introduction Canada's lobster fishery is our most valuable inshore fishery on the Atlantic Coast. In recent years the catch has approximated 50 million pounds with a marketed value of 20 million dollars. Almost 20 thousand fishermen fish about two million traps valued at over five million dollars. In the major producing areas inshore fishermen derive a high proportion of their fisheries income from this one species. Everyone is agreed that conservation of this valuable resource is of primary importance but just what is conservation and what are its objectives? Conservation has been defined as "Wise Use" and a commonly stated objective is to obtain the maximum sustained yield. In marine fisheries the yield is usually thought of in pounds. With staple foods, sufficient production to maintain national nutritional standards is, of course, of primary importance. The lobster, however, is a luxury item that con- tributes relatively little to our nutritional welfare. Even if our total production were consumed in Canada the per capita consumption of edible meat would be less than one pound per year. The per capita consumption in North America is less than two ounces a year. With such a product it can be questioned whether the maximum sustained yield is as important an objective as the welfare of the 20,000 primary producers.

Regulations Over the 100 year history of the lobster fishery numerous regulations have been adopted. These have become so complex in detail as to require a special office con- solidation for ready reference. Of these regulations the three most important appear to be those that (a) prohibit the sale of egg-bearing lobsters, (b) establish closed fishing seasons and (c) provide for a minimum legal size. To what extent have these regulations affected the yield from the fishery or the welfare of the fishermen?

Protection of egg-bearing lobsters Lobsters mature when 7 to 12 inches long and the females spawn every other year. The eggs remain fastened to the abdominal swimmerets until they hatch a year later. Regulations prohibiting the sale of egg-bearing lobsters are in general supported by fishermen, particularly in areas where such lobsters are scarce. Observations in the southern Gulf of St. Lawrence over the past nine years on the newly-hatched, free-swimming larvae show that the hatch varies as much as 2:1. Survival during the one to two month free-swimming period is, however, much more variable-the numbers surviving to the last free-swimming stage varying as much as 13 14

40:1. To date it has not been possible in this area to demonstrate any clear relationship between the number hatched and the number that survive the free-swimming period. In this area lobsters mature early with the result that egg-bearing females and larvae are abundant-apparently many times as abundant as in southern Nova Scotia, an area of late maturity where very few lobsters escape the fishery to become mature. In spite of this great disparity in the abundance of mature lobsters and larvae, the commercial production in the two regions is very similar-about 10,000 pounds per square nautical mile. These observations have failed to provide any factual basis for protecting egg-bearing females. Although the regulation may be of little value in improving the yield or the welfare of the fishermen it seems unlikely that it is actually harmful. n The regulation is popular with most fishermen, it affects a relatively small proportion of the catch, and the lobsters released become legal when the eggs hatch.

Closed seasons Closed seasons were originally introduced in an effort to reduce the fishing intensity and so arrest a steady decline in landings during the early 1900's. Later, adjustments in the seasons were made to aid marketing and to permit fishermen to engage in other fisheries or other seasonal occupations. At present in the Bay of Fundy and southern Nova Scotia the season is open for six to 72 months but along the remainder of the coast for only two to three months. Studies of tag returns, size compositions, seasonal declines in catch per unit effort, days fished, gear set and hauled show that the rate of exploitation is often higher in the short seasons. The fishery has simply adjusted to the shorter season by employing more men, boats and gear. It seems obvious that closed seasons have failed in their primary objective of improving the sustained yield. They have, however, had numerous secondary effects-some good, some bad. With short seasons fishermen can turn to other fisheries, farming, lumbering etc., and in some areas it is only by so doing that they are able to subsist. Closed seasons have improved the quality of the lobsters in that less than 20 per cent are caught from July to September when they are soft-shelled and slack meated, difficult to hold and ship and give a poor meat yield. Closed seasons are difficult and costly to enforce and supplementary regulations restricting fishermen, boats and gear to one fishing season a year have been adopted. These supplementary regulations have been seriously questioned on sociological and economic grounds. With short seasons persons who I have been gainfully employed elsewhere find it advantageous to engage in the fishery when lobsters are unusually abundant. This effectively reduces the catch of the I steady fishermen who find it difficult to see the value of restrictive conservation measures when the accruing benefits are divided among individuals who have not contributed. With year round fishing over half our lobsters would be landed from July to September and would depress an already low summer market resulting from heavy United States production. One possible solution is a universal closed season from July to September to avoid the period of heavy United States production and to 15

limit the sale of newly-moulted lobsters. With such a season weather and ice conditions would permit from four to nine months' fishing in the various districts. Land- ings would be concentrated in the early fall.

Size limits Minimum legal sizes are established with the objective of increasing the sustained yield, either by allowing more animals to mature and so increasing reproduction or by taking advantage of the period where growth is rapid enough to more than offset losses through natural mortality. There is so little evidence of a relationship between I the abundance of mature lobsters and commercial production that the prospects of size limits improving the yield through their effect on reproduction do not appear good. There is, however, considerable evidence that sub-legal lobsters grow enough to more than compensate for natural losses. Extensive marking experiments have shown that lobsters near present size limits survive well and as a general rule moult once a year and grow about 50 per cent in weight. In certain warm water areas growth of the smaller catchable lobsters exceeds 100 per cent a year. Size distribution studies show exceptionally high total mortality rates. Extremely high tag returns (general average of 60 per cent) and the rapid seasonal drop in catch per unit effort show that most of the mortality can be ascribed to the very intensive fishery. These observations provide strong support for minimum size limits. There is also the fact that wherever size limits have been observed the commercial catch has improved. On the basis of the available evidence it is concluded that size limits are the most effective means of increasing the yield from the fishery. If this conclusion is substantiated by controlled experiments, work should be extended to determine the best minimum size limits for the major stocks of lobsters.

Economic and sociological effects

There are two market categories for lobsters, the smaller "canners" which are hermetically sealed or processed as chilled meat and the larger "markets" which are sold alive. There is usually a marked price differential with the markets worth considerably more per pound to the fishermen. From the fishermen's point of view the maximum sustained yield should be measured in value rather than weight. There is, of course, the contribution that the actual processing of canner lobsters makes to the communities. Is this contribution great enough to compensate for the lower price to the primary producer? In the final analysis conservation is action by individuals. It is the individual fisherman who releases undersized or berried lobsters. To be effective, of course, conservation must be practised by all or a large majority of those using the resource. The individual or group of individuals is led to believe and perhaps has a right to believe that he will benefit. In the lobster fishery this is not always the case. In 1947 a group of fishermen at Fourchu, Nova Scotia, agreed to a regulation which required them to release their smaller lobsters to the extent of approximately half their catch. In 1948 they further agreed to limit their traps in the firm belief that they were wasting money in excess effort. The value of the catch increased rapidly to the point where new fishermen were attracted. By 1956 the fleet had increased 16

80 per cent and the number of traps doubled. The catch per boat has started to decline. At Tignish, Prince Edward Island, fishermen were forced in 1954 to observe an unpopular size limit. Catches have improved and in 1956 the catch per boat was the highest in recent history. Judging from 13 years' records at this port it seems certain that the fleet will increase to the point where the average catch per boat is about the same as previously. The rather unique features of the lobster fishery-a readily available but limited supply, a heavy demand and high price, and a relatively low capital investment lead to a degree of exploitation and competition that is perhaps unequalled in marine fisheries. Long-term, enthusiastic co-operation in restrictive conservation measures seems unlikely unless fishermen receive more assurance that as individuals they will benefit. Limited licensing is a possible but unpopular solution. Less drastic measures such as increased licence fees, licence cancellation for infraction of regulations, residence requirements and restriction of government loans for lobster boats and gear might ease the situation considerably. Any reduction in boats or gear should increase the net value of the catch. Possibly our primary objectives in the management of the lobster fishery are (1) discover and apply the minimum size limits that will yield the greatest gross value and (2) reduce the fishing intensity to increase the net value of the catch and give the fishermen who practise conservation greater assurance that they will benefit. The problem involves a complicated interplay of biological, economic and sociological factors and is of sufficient importance to warrant detailed study. Some Sociological Effects of Quota Control of Fisheries by J. L. Hart, Director

Fisheries Research Board of Canada, Biologica! Station, St. Andrews, N.B. Recently striking changes have occurred in many of our fisheries. Some of them seem to centre around the management of the fisheries by quota. Most of these changes, when not forced upon us by circumstances, have been economically advantageous. They have, however, strongly affected in one way or another the lives of the people engaged in the fishery. Superficially some of the results do not appear to be improvements. If the well-being of the people is in fact impaired by quota control, it may be appropriate to examine the sociological implications of this tool of management. The following comments point out some of the problems. Some of these may seem to justify investigation by sociologists. Quota management of fisheries is resorted to in cases where we presumably know how much fish can be removed from a stock without diminishing the contribution of the stock to our economy. It is applied by allowing fishing from an opening date until the amount of the pre-determined quota is caught up. Apparently where we have the required biological information, quotas provide a seemingly logical method of management for maintaining fish stocks at levels where their yield is maximum or otherwise optimum. Any type of management will have economic and sociological ramifications. Catch quotas are no exceptions. And as in the cases of other means of management economic and sociological effects are inextricably mixed. My interest in the sociological results of quota regulation was first roused at a public hearing on halibut regulations for the International Fisheries Commission. As background, I might mention that quota regulation of halibut had been followed by a great increase in availability of fish which led to very fast fishing. The good fishing attracted additional fishing efforts. As a result of the high reward for fishing effort, and of the increased effort, the quotas were (and are still) caught up very quickly. The incident involved the protest of a veteran Norwegian skipper. I wish I could accurately reproduce in full his dignified but perturbed statement. Instead I shall have to provide a poor paraphrase-something like this: "They say that the halibut is being conserved and perhaps it is. I don't know. I do know that before the halibut was conserved and I worked on the schooners I used to fish halibut nine months of the year and make a decent living. Now the season is so short a boat can't catch enough fish to provide a year's income. So I have to turn to working on salmon packers and other things that don't suit me." So much for our Norwegian friend. He said enough to make it clear that he had suffered severe vocational dislocation as a result of management policies imposed to increase the growth of the halibut biomass. For him it was an important matter. It seems that sociological stresses result from quota regulation because the regulation accentuates competition among fishermen on the grounds. In some cases there 17 is

are co-operative aspects to operations within fishing fleets so that for any situation of fish density and distribution there may be an optimum amount of fishing effort. How• ever, fishermen are always in competition with each other. The fish caught and boated by one fisherman are no longer available to his fellows. However, under quota regulation competitive effects are reinforced. The fish landed by one fisherman reduce the chances of his competitors (and of himself, too, for that matter) of catching the fish which remain in the water. The situation could be approximated under some circumstances where fishermen were in competition for markets, but such occurrences are not common in stabilized fisheries. The competition leads to an hectic scramble for fish, inordinately long hard working days, perhaps taking unwise chances, near financial disaster to the individual from breakdowns in gear and equipment or from temporary illness. To me these things do not seem good from the point of view of the fishermen involved. It is more difficult to assess the effects of concentrated production on the economics of fishing. While it seems likely that much uneconomic expense is incurred for competition, there must be real advantage in the mechanization that is encouraged. The situation for packers and secondary handlers of fish and their shore crews is also difficult to assess. To do so one must weigh the advantages of concentrated predictable operating seasons and their evident economies and rewards against the disadvantages of general rush, crowded facilities, and overworked crews. The effects of congestion and delay on the quality of the fish placed before the consumer might also be considered as a sociological result. The main sociological effect of quotas would appear to be the discouragement of specialized full time fishermen or of fishermen who specialize in one particular fishery. The fisherman is forced into other fisheries or other occupations in order to maintain his position in society. To some very worthy people this is a most unwelcome development. Whether it is actually bad or not would appear to be a problem for the sociologists. If bad, it must be decided whether to endure the shortcomings so as to enjoy the advantages of a tidy method of management, or to modify the management method to adapt it to human needs. Another and related effect arises from the very productive fishing each year at the beginning of the quota period. This encourages many fishermen to take part in harvesting the resource. Some of them are part time fishermen or men who are only moderately well qualified for the work. Their participation adds money and variety to their own lives. The fish they catch, however, are taken away from the potential landings of professional fishermen. Some of the foregoing remarks may suggest basic antagonism to quotas as a method of management. There is no such antagonism. Management of a fishery is complicated and none of the applicable tools is perfect. I recognize that quotas also have shortcomings. We must realize also that quota control shows strongly features which are present in some other devices of management. For example, seasons, whether imposed by regulation or by the availability of the fish, when the size of the exploitable stock is limited, may have similar effects. 19

In closing I might add that trial has shown that the headlong exploitation stimul- ated by quotas even has some disadvantages from the point of view of maintaining the fish stock. The halibut fishery-opening intensively at the beginning of the growing season for the fish-derives little advantage from the growth of the current year. In addition intensive fishing for halibut takes the quota before all parts of the stock are fully available and thus failed to make complete use of the resource until special adjustments were invoked. To me it seems that these biological defects in the quota system resemble what I regard as the sociological shortcomings. Each defect arises from ignoring the complexities of its subject matter.

Some Economic Aspects of Control by Quota by W. C. MacKenzie, Director, Economics Service,

Department of Fisheries of Canada. The optimal allocation of resources in the fishing industry has been discussed in several recent papers' and an international gathering2 to discuss the subject has also taken place. In the following notes, an attempt is made to present the gist of these discussions with particular reference to catch quotas as an instrument of fishery management policy. In view of the interests of the present meeting, this question is approached here in its relationship to fishery conservation. Conservation may be defined, in economic terms, as postponement of the use of a resource. Its rationale is similar to that of the investment of capital generally: postponement of utilization is expected to result in a larger supply being available at a future time. As applied to the fisheries, it follows that the optimum degree of utilization must be conceived of as a time-function of some kind, i.e. a catch per unit of time. Achieve- ment of this optimum involves a synthesis or reconciliation of the dynamics of fish populations, the rate of withdrawal from the stock (by fishing) and the time-preference schedule of the community concerned (or the rate of interest on invested capital). How may such a complex objective be realized? Let us consider first of all the nature of the principal instrument for its realization, viz. the primary fishing enterprise. That is, the unit of ownership and management- the two functions are normally combined in the same person or group-operating fishing craft with the accompanying gear and making the first sale of the fish taken. These units range in size from single own-account fishermen to relatively large firms (or divisions of firms) operating fleets of several vessels. Generalization, concerning such things as the shape of supply curves and the variability of labour and capital factors for primary fishing enterprises, is seldom possible-even within a single country, and especially on an international basis. At least one characteristic is exceedingly persistent and widespread, however. That is the sharing, or "lay", system of remuneration. Returns to both capital and labour factors are based on prearranged shares in proceeds from the sale of the catch. There is no return to the resource factor, i.e. rent. We'll return to this in a moment.

'Crutchjield, James, A., "Common Property Resources and Factor Allocation", The Canadian Journal of Economics and Political Science, vol. 22, no. 3. Gordon, H. Scott, "An Economic Approach to the Optimum Utilization of Fishery Resources", The Journal of the Fisheries Research Board of Canada, vol. 10, no. 7; "The Economic Theory of a Common- PRopesty Resource: the Fishery", The Journal of Political Economy. vol. LXII, no. 2. Scott, Anthony, "The Fishery: the Objectives of Sole Ownership", ibid., vol. LXIII, no. 2. i`ïhe Round `I-able on Fisheries, under the sponsorship of the International Economic Association, con- vened at Rome, Italy, Sept. 13-18, 1956. 21 22

The ubiquity of the share system is related, apparently, to tne ract tnat it permits the peculiar risks incurred in fishing operations being shifted or spread to some extent. For that reason, the system is thought in some quarters to encourage investment- in expansion or innovation, for example. On the other hand, because of its comparative rigidity, it is sometimes held to have the opposite effect. On the whole, its significance in this respect does not seem to be great-at least in the Canadian milieu, where renegotiation of shares (to facilitate adaptation to technological change) may be accomplished usually without difficulty. Of very great significance, however, is the point we touched upon a moment ago: the absence of a rental return in the fishing industry. This arises out of the fact that fisheries, in contrast with most other sectors of the economy, are exploited generally under conditions of tenure which make the resource common property. Being unowned -indeed impossible of ownership, probably, in most cases-the resource cannot yield a rent. This may be explained as follows. Let us suppose that, in the case of a particular fishery, supplies have been drawn from the more accessible and/or richer grounds at relatively constant costs. At some point, as the demand for the products of the fishery grows, operations will be intensified beyond the constant-cost level, or extended to more distant and/or less rich grounds, or both-resulting in higher costs at either the intensive or extensive margins. Under these conditions, if, as in the case of land resources, for example, the grounds were possible of ownership, an income increment (known as rent) would accrue to the owners of the more accessible and richer ones. Now, since total cost tends to equal total revenue (by definition, as it were), the absence of rent (explicit or implicit) results in a gap that may be filled by an increase in capital costs. In short, the industry probably has a "built in" tendency toward over-capitalization 3 Over-capitalization represents a misallocation or mal-allocation of production factors. The effect is to dissipate, in excess capacity' (numbers of fishing craft with their complement, etc.), any increase in aggregate returns to the industry. What are the implications of all this for control policy? The above outline probably simplifies drastically-not to say vulgarizes-^vents in the real world. Nevertheless, insofar as it does approximate reality, it has very important implications for our problem. Assuming, as we may in normal circumstances, a continuous secular growth in demand, the imposition of restrictions on the catch, i.e. a quota or the like, designed to maintain the fish stocks at or restore them to a predetermined level, will be followed by a rise in price and consequently in aggregate returns. The gap mentioned earlier will be widened and the tendency toward development of excess capacity will be strengthened. In other words, the quota will be filled with successively larger numbers of vessels, each taking fewer fish in a shorter period of time. The effect is not only to waste resources in over-capacity in the case of the fishery under quota. The development of a fleet capable of participation in other fisheries

This applies to the industry as a whole and, as shown later, such a tendency is compatible with under • capitalization in the units of the industry, i.e. in individual fishing enterprises. The stand-by capacity required to provide for variation in raw-material supply and the flow of distribution is, of course, a different matter altogether. 23 during the "off season" will be fostered. This tends to generalize over-capacity by forcing enterprises in the latter fisheries to diversify operations and equipment also.5 The impairment of efficiency, from the economic and social point of view, extends beyond the primary fishing industry and into the secondary and tertiary (or trade) phases of the industry. The shortening of the operating season increases overhead costs (per unit of output) in handling and processing. In addition to higher direct costs, the prolonged storage of fish products-required to adjust concentration in the supply period to the comparatively regular flow of distribution-results often in serious loss of quality. The extent to which these tendencies are realized in practice may be illustrated from the recent history of the Pacific halibut fishery-perhaps the outstanding example of a fishery under quota control. During the past 25 years or so, while the catch of Pacific halibut increased by about 20-25 per cent, the number of vessels engaged in the fishery increased by some 120-125 per cent. The figures understate the increase in fishing intensity, since no allowance is made for the growth in productivity resulting from improvements in design, construction, power and equipment with electronic fish-finding and communication devices. At the same time, the fishing season has been reduced to less than ten per cent of its original length on one of the principal grounds and to about twenty per cent on another. As a concomitant of this, over eighty per cent of the catch is now frozen as compared with about forty-five per cent at the beginning of the period mentioned.e It may be said too that, according to estimate, about one-third of the Pacific halibut fleet is idle between seasons. Some proportion of the vessels that do obtain employment, doubtless, would be under-utilized in operations for which they are imperfectly fitted. A quota or similar limitation on the catch, below the level that would be forth' coming on a free market, must also have an important influence on price formation at waterfront or dockside markets-through its effect on the supply function of the primary producers. In the short run at least, supply will be inelastic as to price. A reduction in price will tend to result in a longer season, rather than a shorter supply -assuming our earlier analysis to have some validity. In fact, the supply function would become elastic only if prices fell to the point where the attraction of occupational alternatives reduced fleet capacity below the level at which the quota could be filled by the remaining vessels fishing on a full-time basis. The further implications of this kind of supply situation would depend on the organization or structure of the dockside markets. In perfectly competitive markets, buyers' margins would be minimized and the fullest development of over-capacity in primary production would be encouraged. Completely monopsonistic markets7 would lead to opposite results: excessive costs (due to the existence of over-capacity) would be minimized, but at the expense of equitable distribution of income. The actual position, of course, lies somewhere between these extremes.

bUnder certain conditions, of course, overall efficiency may best be served by the development of flexibility in fleet operations-as, for example, when seasonality or relative price changes make some mobility among different fisheries desirable-but that does not affect the principle here. 6Technological progress in fish processing and marketing has contributed to this development, of course. 7That is, markets dominated by a single buyer. 24

Enough has been said to indicate that quota control may involve substantial diseconomies, both internal and external, for the fishing industry. Can anything be done about it? It is probable that a program restricting the catch to a level consistent with, say, maximum sustainable physical yield would not deviate seriously from the social optimum if inputs could also be controlled. Such control would mean restriction on entry to the fishery or fisheries involved. To accomplish this, since, in most cases, private ownership of fishery resources is out of the question, it has been suggested that the state assume a property right in these resources and collect a rental for the privilege of access to their use. The most efficient fishing enterprises would be able to offer the most rent (in the form of a share of the catch or of an absolute sum of money) and the least efficient enterprises would be eliminated from the industry. With perfect competition, all the rent would accrue to the state and efficiency would be maximized. Under less perfect conditions, the number of privileges would have to be limited more arbitrarily on the basis of some calculation of the (fully-utilized) capacity needed to fill the quota within the period in which fishing operations normally are possible. In any event, competitive bidding for fishing privileges would be necessary at intervals to facilitate the movement of capital and labour into and out of the industry. A market in such privileges might arise, under certain conditions, as enterprises changed hands or engaged in different combinations of activity. If there were no need periodically to change the number of privileges, this market could be permitted to allocate them. As to what should be done with the rents collected, these might accrue to the state as landlord, or, at the outset, they might be used to resettle or compensate displaced fishermen, or, later on, they might be used for purposes like research and protection for the controlled fishery or fisheries. Redistribution to the privileged fisher• men, another possibility, would probably result in their being discounted in the application for privileges. Despite the advantages of this scheme, there are formidable obstacles in the way of its implementation. We need not be too greatly concerned, perhaps, with the strengthening of the fishermen's bargaining power vis-a-vis the fish buyers. Nor need consumers be adversely affected, if, as seems likely, the ability of a producers' monopoly to raise fish prices is limited by cross-elasticities of demand or substitutability among the group of animal protein food products. Producers excluded from the fishery or fisheries, and thus forced to liquidate investments in unfavourable circumstances, could be compensated as already suggested. Insuperable difliculties, however, might arise in the case of international fisheries. Since total production costs include the "opportunity" incomes of fishermen, countries with a low level of living would tend to pursue a more intensive fishery than others to achieve an optimum degree of exploitation of a given resource (fish stock)-the optimum being defined as the maximum net economic yield or, in simplified terms, as the maximum difference between total production costs and total value of catch. There would be no one unique optimum level of fishing intensity, therefore, for a resource that is subject to international exploitation-there might be several. If so, agreement among the countries concerned to a system of control like the one described here would probably be precluded. CORRECTION

In Issue 21 of "The Canadian Fish Culturist," published December, 1957, the following corrections should be noted:

Under "Literature Cited"-page 17, line 2, for "000-000" read "19-23", line 4, for "000-000" read "1-6"; page 23, line 2, for "Issue VI, pp. 00-00" read "Issue 21, pp. 7,17", line 4, for "Issue VI, pp. 00-00" read "Issue 21, pp. 1-6", line 6, for "No. 62, pp. 17-2" read "No. 62, pp. 17-23"; page 31, line 7, for "No. VI, pp. 00-00" read "No. 21, pp. 7-17", lines 8-9, for "No. VI, pp. 000-000" read "No. 21, pp. 19-23".

On page 20, Table I, headings for the last two columns should read:

Survival from

Under Eggs yearlings

100C 6% 54',-o 4ô

ISSUE TWENTY-THREE DECEMBER - 1958

THE CANADIAN FISH CULTURIST

vBR.^R,Y ^^t^I^^RIES AND O CFAn

Published at Ottawa by The Department of Fisheries of Canada ISSUE TWENTY-THREE DECEMBER - 1958

THE CANADIAN FISH CULTURIST

Published at Ottawa by The Department of Fisheries of Canada

THE QUEEN'S PRINTER AND CONTROLLER OF STATIONERY OTTAWA. 1958 b941?-6-1 CONTENTS

Page Observations on the Spawning of Lake Trout, Salvelinus namaycush, and the Post-Spawning Movement of Adult Trout in Lake Simcoe - H. R. McCRIMMON...... 3

The Survival of Yearling Lake Trout Planted in South Bay, Lake Huron - F. E. J. FRY and J. C. BUDD ...... 13

Notes on the Food of the Young of Three Species of Pacific Salmon in the Sea - MURVEL E. ANNAN ...... 23

A Direct-Current Electrofishing Apparatus Using Separate Excitation - A. R. MURRAY ...... 27

Back-pack Fish Shocker - A. A. BLAIR ...... 33

Experiments with Toxaphene as a Fish Poison - GEORGE E. STRINGER and ROBERT G. McMYNN ...... 39 Review of Literature: "Freshwater Fishery Biology," by Karl F. Lagler - Re- viewed by VADIM D. VLADYKOV ...... 49

The Canadian Fish Culturist is published under the authority of the Minister by the Department of Fisheries of Canada as a means of providing a forum for free expression of opinion on Canadian fish culture. In the areas of fact and opinion alike, the responsibility for statements made in articles or letters rests entirely with the writers. Publication of any particular material does not necessarily imply that the Department shares the views expressed. In issuing The Canadian Fish Culturist the Department of Fisheries is acting only as an instrument for assisting in the circulation of information and opinion among people in the fish culture field. Those who may wish to discuss articles which have been published in The Canadian Fish Culturist are encouraged to do so and space will be made available.

Correspondence should be addressed to the DIRECTOR, INFORMATION AND EDUCA• TIONAL SERVICE, DEPARTMENT OF FISHERIES, OTTAWA, CANADA.

Published under Authority

HON. J. ANGUS MACLEAN, M.P., Minister of Fisheries Observations on the Spawning of Lake Trout, Salvelinus Namaycush, and the Post-spawning Movement of Adult Trout in Lake Simcoe by

H. R. McCrimmon

Ontario Agricultural College, Guelph, Ont.

Introduction

The material presented in this paper comprises a summary of observations made on spawning populations of lake trout in Lake Simcoe between 1951 and 1957. The majority of the data was collected at the extensive Glenrest Beach shoals at the north end of the lake where submarine trap-nets were set each autumn by the Fish and Wildlife Division of the Ontario Department of Lands and Forests for the collection of lake trout eggs as a part of Ontario's fish culture programme. Records of the post- spawning movements of parent fish from the Glenrest Beach shoals were obtained by the recovery of tagged fish by anglers.

Description of Lake Simcoe

Lake Simcoe, the fourth largest inland lake of Ontario, has an area of 280 square miles but a relatively short and generally exposed shoreline of only 144 miles, including that of the few islands. The average depth has been calculated at 56 feet with some six per cent of the lake over 90 feet deep (Rawson, 1930). The main body of the lake is oval in shape with a long deep bay, Kempfenfeldt Bay, on the west and a shallow bay, Cook's Bay, on the south. The shoreline of the lake is about 55 per cent stony, 35 per cent sandy, and the remaining 10 per cent capable of supporting vegetation in rich muddy bays. On account of the wide variety of habitat conditions present, Lake Simcoe is ideally suited for the production of both warm water and cold water fishes which enables an active sport fishery at all seasons of the year (McCrimmon, 1956). Habitat conditions are very favourable for the reproduction, growth, and harvest of lake trout.

3 59417-6-2 4 Spawning Grounds

Lake Simcoe provides extensive shoal areas for the spawning of lake trout. The trout spawn on shoals located along the shores of the mainland and few islands, also on several submerged mid-lake shoals. These shoals are covered characteristically by wave-washed gravel, rubble, and stones, and are quite free of silt. Some 50 per cent of the shoreline of the lake is suitable for trout spawning and at some locations the stony beaches extend off-shore in excess of 500 feet before reaching water depths of 20 feet. Several submerged shoals are over a mile in length. Spawning generally takes place in less than 20 feet of water, occurring frequently at depths of two to ten feet. Pre-Spawning Movement of Adult Trout

The lake trout appeared at the spawning shoals of Lake Simcoe early in October, the actual dates varying between October 4 and October 9 in the years from 1951 to 1956. The autumn migration of the trout from the depths of the lake was, of course, coincident with dropping surface water temperatures and the fall turnover. During the years of study, surface water temperatures were between 50° and 57°F. at the time of arrival of the lake trout. The arrival of the trout on the spawning shoals each year followed one or more.days of strong winds, averaging over 10 m.p.h. in velocity for a 24-hour period and with gusts in excess of 15 m.p.h. These statistics support the observations of local fishermen that strong winds are necessary before the trout come to the shoals.

A summary of weather conditions on the day previous to the evening when trout were first observed on the shoals is given in Table I.

Table I

Summary of Weather Conditions on First Date of Appearance of Lake Trout on the Shoals

Air Temp. Wind Conditions Surface Hours Water Deg. F. M.P.H. of Year Synopsis Temp. Sun- Deg. F. Max. light Min. Max. Av. (1 hr.)

1951 52 44 50 Strong 20 27SW 0.0 Overcast, drizzle rain- showers

1952 50 45 55 Mod. 11 20SW 2.5 Overcast, with clear intervals

1953 54 54 75 Mod. 15 19N 2.9 Generally overcast, rain- showers

1954 54 32 50 Mod. 11 23NW 3.2 Overcast by 10 a.m. rain

1955 57 57 61 Strong 19 25E 0.0 Overcast and rain

1956 57 41 55 Mod. 12 26NW 8.5 Cloudy, clearing by 9 a.m. 5 Activity of Trout on Spawning Shoals

The time during which the adult lake trout are concentrated on the spawning shoals can be divided conveniently into two periods: the pre-spawning interval when the trout frequent the shoals but are not yet ready to spawn, and the spawning period when the eggs are laid and fertilized. Chronological data on the movements of trout to the shoals and the subsequent spawning periods are given in Table II.

Table II Observations on the Autumn Movement and Spawning of the Lake Trout

Night of Arrival Night of First Year Peak Spawning Night of Last Duration of of Trout at Shoals Known Spawning Period Known Spawning Spawning Period

1951 Oct. 8 Oct. 17 Oct. 21-25 Oct. 29 12 days

1952 Oct. 5 Oct. 15 Oct. 17-19 Oct. 24 9 days

1953 Oct. 6 Oct. 16 Oct. 19-21 Oct. 25 9 days

1954 Oct. 6 Oct. 21 Oct. 25-28 Nov. 1 9 days

1955 Oct. 5 Oct. 24 Oct. 27-30 Nov. 2 10 days

1956 Oct. 4 Oct. 14 Oct. 2 Oct. 29 16 days

(1) Pre-Spawning Intenal The adult lake trout frequented the spawning shoals of the lake for a variable length of time each year before actual spawning began. This period between the advent to the spawning shoals and the initiation of spawning each year varied from a minimum of nine days in 1951 to a maximum of 19 days in 1955. A summary of weather conditions which occurred during the pre-spawning interval is tabulated in Table III.

Table III

Summary of Weather Conditions between the Arrival of Trout on Shoals and Initiation of Spawning Part (a)

Duration Av. Water Water Temp. Average Mean Year of Temp. Deg. F. Total Degrees Air Temperature Interval at Surface Below 60°F. Deg. F.

1951 9 days 53 (50-60) 61 51 (38-72) normal

1952 10 days 52 (48-56) 48 (32-66) below normal

1953 10 days 55 (52-61) 50 (30-69) below normal

1954 15 days 56 (51-58) 54 (28-74) above normal

1955 19 days 55(50-66) 62 (30-75) above normal

1954 10 days 54 (54-57) 50 (31-72) below normal

59417-6-2; 6

Part (b) Table III.-Continued

Wind Conditions Total Hours Year of °Jo of Days with Synopsis Sunlight Av. Daily (m.p.h.) Max. Vel. Greater Than 10 m.p.h.

1951 60.6 7.4 13 Generally clear, no rain 1952 61.5 8.1 0 Generally clear, rain only last 2 days 1953 64.8 8.6 10 Generally clear, rain only on Oct. 11th 1954 64.8 12.7 50 50% clear, 50% overcast, rain on 6 days in mid-period 1955 60.8 9.9 62 40% clear, balance rain

1956 61.6 10.7 60 Generally clear, rain only on Oct. 7th and 9th

A comparison of a number of measurable weather conditions (see Tables III and IV) occurring during the pre-spawning interval which may have influenced the length of the interval suggested a correlation only with the total hours of sunlight accumulated after the arrival of the trout on the shoals. In each year approximately 60-65 hours of sunlight had occurred at the time of first spawning. While it is recognized that spawning began each year with water temperatures between 52° and 57°F., temperatures in this range did not precipitate spawning. In five of the six years between 1951 and 1956, water temperatures had dropped to 55°F. or less by October 9 but spawning did not begin until some days later when some 60 hours of sunlight had been accumulated. No correlation was apparent between the duration of the pre- spawning interval and other limnological, meteorological, or astronomical factors studied. The significance of the apparent correlation between accumulated sunlight following the arrival of the trout on the shoals and the initiation of spawning, which may be only a coincidence, would seem worthy of further investigation.

(2) The Spawning Period A considerable variation has been observed in the date of first spawning during the years of study, a maximum difference of 10 days being noted between the start of spawning in 1955 and 1956. Table IV Summary of Weather Conditions on the First Day of Spawning.

Night of First Surface Wind Duration of Phase Year Known Spawning Water Conditions Sunlight of October Temperature During Day Hours Moon

1951 17th 53°F light (6.3NE) 6.4 past full 1952 15th 52°F light (8.2SW) 1.0 none 1953 16th 56°F light (5.4NW) 9.0 past 1 Q. 1954 21st 52°F light (7.5NW) 9.7 third Q. 1955 24th 52°F strong (19SW) 0.5 past I Q. 1956 14th 54°F light (5.9SW) 7.4 nast I Q. 7

Reference to Table IV reveals no uniformity in weather conditions at the initiation of the spawning period other than a water temperature of 52° to 56°F. Winds were generally light in contrast to the strong winds that marked the arrival of the trout on the spawning beds although heavy winds occurred during the first day of spawning in 1955. Sunlight varied from a maximum in excess of nine hours to overcast skies. Rain occurred on two occasions. As noted above in the text, it is possible that the extent of accumulated sunlight during the pre-spawning interval may directly or indirectly influence the time of spawning.

The spawning period extended over a period varying from nine to 16 days between 1951 and 1956. A summary of average weather conditions which prevailed during each of the spawning periods is given in Table V.

Table V

Summary of Weather Conditions During the Lake Trout Spawning Period

Water Temperature Year Duration Air Temp. Hours of in Days Average Day Degrees Deg. F. Sunlight Deg. F. Below 60°F.

1951 12 59 70 1952 9 49 58

1953 9 55 30

1954 9 52 56

1955 10 49 97

1956 16 52 119 55 84

Observations on the spawning activity of the lake trout revealed that maximum activity was usually reached within four days after spawning began and continued at a high level for the next three to five days, by which time the majority of trout were spawned out. The spawning activity then decreased abruptly and there was an exodus of trout from the spawning areas within the next three or four days.

An analysis of environmental factors which might be expected to affect spawning activity has shown no definite correlation of any factor other than wind. Local fishermen state that heavy onshore winds and rough water generally result in a short spawning season whereas calm weather allows spawning over a more prolonged period. Observations made on the spawning runs at Glenrest beach indicated that strong onshore winds hastened the peak in spawning activity once the fish had begun to spawn and this shortened the overall spawning period. A comparison of the more-or- less normal spawning periods of the years between 1951 and 1955 with the extended spawning period of 1956 shows a relationship between onshore winds and spawning activity. 8

In the years from 1951 to 1955, onshore winds followed closely after the initiation of spawning and peak spawning activity occurred within two to four days. In contrast, during the extended spawning season of 1956 there were no onshore winds until over a week after spawning was first noted and spawning activity remained at a low level. Spawning activity increased rapidly in 1956, eight days after spawning had begun with the arrival of moderate southerly onshore winds which developed into strong winds for the remainder of the peak spawning period.

Species of Fish Frequenting Lake Trout Spawning Shoals The captures of fish made by two 15-foot submarine trap-nets on the Glenrest beach spawning shoals during October for the collection of lake trout eggs are recorded in Table VI. The kinds and numbers of fish frequenting the shoals are apparent in this Table. Between 1951 and 1956, some 3,547,000 fertilized lake trout eggs were collected for culture. Table VI

October Capture of Fish by Two Submarine Trap-nets Set at the Glenrest Beach Spawning Shoal between 1951 and 1956.

Number of Fish Captured

É m

^0 Duration of i Netting 0 ^ ô w 3 ^

E .T 0 tlJ ÿ F_ d Ÿ E ti Cj .^7 vl -0 ul y O o .

â ^ O. '>^ Û ci! , v, Û rn o Û

Year Operation

1951 Oct. 3 to Oct. 30 1341 2004 435 0 232 2 23 63 8 193 20 0 1

1952 Oct. 7 to Oct. 25 1218 2987 126 0 952 0 40 10 11 522 86 0 0

1953 Oct. 5 to Oct. 23 835 425 1368 0 735 0 33 50 11 228 3 4 0

1954 Oct. 5 to Oct. 30 1089 4130 352 1 1551 1 16 23 11 710 16 2 0

955 Oct. 3 to Nov. 2 1135 7694 238 0 674 0 26 206 2 465 68 0 0

956 Oct. 5 to Oct. 30 445 10588 2535 0 269 0 22 65 4 531 48 0 17 ------Total...... 6063 27828 5054 1 4413 3 160 417 47 2649 241 6 18 9 Post-Spawning Movement of Adult Lake Trout

A total of 500 lake trout were tagged from captures of 3,059 trout made by submarine trap nets set on the Glenrest Beach spawning shoals at the north end of the lake during the early part of October in the years 1953, 1954, and 1955. The trout were tagged with coloured plastic tags attached at the base of the dorsal fin by nylon thread. Most records of the recovery of tagged fish were provided voluntarily by anglers. The purpose of the study was to follow the movements and distribution of adult lake trout after leaving the spawning beds in order to determine firstly, if the spawning population was a local one restricted to a segment of the lake and, secondly, if there was any evidence of a homing behaviour among the trout. A summary of the recovery of tagged trout is given in Table VII. The locations of recoveries are shown in Figure 1.

Figure 1.-Raovtries of tagged take trout in Lake Simcoe, G = Location of tagging, S= Summer recoveries, W = Winter recoveries. 10

Table VII

Summary of Recovery Records for Adult Lake Trout Tagged on Glenrest Beach Spawning Grounds.

Year of Number Maximum Tagging Time of Recovery of Trout Lineal Distance Percentage Recovered Travelled Recovery

1953 Winter 1953-54 ...... 8 8.3 mi. Summer 1954 ...... 7 10 . 1 (150 fish Winter 1954-55 ...... 2 12.8 tagged) Summer 1955 ...... 3 9.0 Winter 1955-56 ...... 7 11.2 Summer 1956 ...... 2 8.5 Winter 1956-57 ...... 1 5.3

Total ...... 30 20.0

1954 Winter 1954•55 ...... 2 8.2 Summer 1955 ...... 4 10 . 5 (200 fish Winter 1955-56 ...... 1 8 . 2 tagged) Summer 1956 ...... 0 - Winter 1956-57 ...... 5 9.8

Total ...... 12 6.0

1955 Winter 1955-56 ...... 6 8.0 Summer 1956 ...... 3 11 . 3 (150 fish Winter 1956-57 ...... 2 3 . 8 tagged) _ Total ...... 11 7.3

Grand Total...... 53 12.8 10.6

A study of the recapture records shows (1) that 10.6 per cent of the 500 lake trout tagged between 1953 and 1955 had been recovered by March 1, 1957. Additional recoveries are anticipated. A comparatively high return of 20.0 per cent resulted from the 1953 tagging, a low figure of 6.0 per cent from the 1954 tagging, and a return to date of 7.3 per cent for the 1955 tagging project. Over 63 per cent of the recoveries were made by winter anglers, 37 per cent by summer anglers, and the remainder by trap nets set at Glenrest Beach.

(2) That those trout which spawned at the Glenrest Beach shoals became widely distributed in Lake Simcoe at other seasons of the year and occupied all those deep basins of the lake where trout normally congregate. The greatest distance recorded between the point of release and that of recapture was nearly 13 lineal miles. As the lake trout are limited to several deep-water areas during the heat of summer and recaptures of tagged trout occurred at each of these locations, one may conclude that there was a mixing of trout from all spawning areas of the lake during the summer stagnation period at least. 11

(3) That there was no evidence of a homing behaviour to the spawning beds among the lake trout, only two of the tagged fish having re-appeared in the trap nets set on the Glenrest Beach shoals although nearly 2,7001ake trout were handled during the 1954, 1955 and 1956 spawning seasons. The homing behaviour recorded by several authors (Martin 1957) may be influenced in Lake Simcoe by the unusually great expanse of suitable spawning beds and the similar character of the various shoals. Summary

Observations were made on the spawning and movement of adult lake trout between 1951 and 1957 in Lake Simcoe, a lake ideally suited for the production of this species and its harvest by anglers. The adult lake trout approached the spawning shoals during early October in each of the study years when surface water temperatures had dropped to between 52°F. and 57°F. Strong onshore winds occurred immediately before the arrival of the trout and would appear necessary to bring the trout to the shoals. The lake trout frequented the shoals for a pre-spawning interval between nine and 19 days before actual spawning began. A relationship observed between the duration of this period and the extent of accumulated sunlight would seem worthy of further study. The spawning period varied from nine to 16 days. Maximum spawning activity was reached within four days (except in 1956), continued at a high level for from three to five days, then decreased rapidly with an exodus of trout from the shoals within the next three or four days. Strong onshore winds appeared to hasten the peak in spawning activity and thus shorten the overall spawning period. Tagged lake trout from a specific spawning shoal became widely distributed in the lake at other seasons of the year, the recovery of over 10 per cent of these trout representing a mixing of trout from all spawning areas of the lake at least during the summer stagnation period. There was no evidence of a homing behaviour among the adult lake trout in Lake Simcoe. References MARTIN, N. V. 1957. The reproduction of the lake trout in certain Algonquin Park, Ontario lakes. Manuscript, Ontario Department of Lands and Forests.

McCRIMMON, H. R. 1956. Fishing in Lake Simcoe, Pub. of Department of Lands and Forests, Prov. of Ontario. RAWSON, D. S. 1930. The bottom fauna of Lake Simcoe, Univ. of Toronto Studies, Pub. Ont. Fish. Res. Lab. No. 40.

59417-6-3

The Survival of Yearling Lake Trout Planted in South Bay, Lake Huron* by F. E. J. Fry and J. C. Budd

Department of Zoology, University of Toronto, and Research Division, Ontario Department of Lands and Forests.

Introduction Shortly after the Fisheries Laboratory had been established on South Bay, an inlet into Manitoulin Island which lies in northern Lake Huron, it became evident that there had been no substantial lake trout spawning there after 1944 or 1945. The reason for this lack of adult lake trout was not immediately apparent but subsequently it has appeared certain that at least the final reduction and the continued absence of adult trout have been due to predation by the sea lamprey, Petromyzon marinus. Because of the lack of native spawning stock, a programme of experimental planting of marked yearling lake trout was begun in 1949 and plantings were made each year thereafter, except for 1950, ending in 1955. In addition, a spring planting of unmarked fry was made in 1955. The details of these plantings are given in Table I. The main purpose of these plantings was simply to determine as accurately as possible the degree to which such fish would survive to catchable size and to compare their growth rate and behaviour with data already available from native South Bay lake trout. However, with the continued predations of the sea lamprey, the plantings have also enabled us to extend our study of the effect of the lamprey on the lake trout beyond the period when the natural population of this species had disappeared. It is hoped now, in particular, to gather data which may be of value in predicting the success of any future efforts to rehabilitate the lake trout in the upper lakes when the current sea lamprey control programme shows promise of success.

Planting Procedure

The stocks planted were from Lake Superior except for one year when some Lake Simcoe fish were included. They were reared to the end of the second summer in various Provincial Hatcheries. (See Table I.) A few days before planting the fish were moved to the Sandfield rearing station about ten miles from the South Bay laboratory.

•Presented at the 10th meeting of the Canadian Committee for Freshwater Fisheries Research, Ottawa, January 4, 1957.

13 59417-6-31 14

It is felt that the recovery period thus allowed, after a long trip either by truck or by air, is a valuable precaution against planting them in an overfatigued state with lessened chance for survival (Black 1956). They were then moved in small loads to the laboratory and placed in three-foot square hatchery tanks aboard the boat from which they were to be planted. The fish were planted over broken rock along the shore of the upper basin of the bay in 10 to 20 feet of water. The areas chosen for planting were the same areas where native lake trout were known to have spawned. The fish were distributed evenly along the shoreline by tossing them overboard from the slowly moving boat by means of a scap net. Only from three to five fish per netful were handled, thus allowing the planter to keep an accurate count. Distribution was at the rate of approximately 2,000 fish per mile of shoreline. The plantings were delayed until surface temperatures were 55°F. (12.8°C.) or lower. The plant of unmarked advanced fry made in May of 1955 was air-dropped in the manner commonly used in stocking inaccessible lakes. Some weeks prior to planting, all yearlings were marked by removing the adipose fin with a razor blade. The same mark was used on every plant, reliance being placed on scale reading to separate the groups when captured. Approximately 1,400 planted fish, so recognized by the absence of their adipose fins, have been recovered to date. Along with these, 185 other lake trout have been taken for which the lack of an adipose has not been recorded or for which its presence has been specifically noted. The fish thus recognized as unmarked have been consistently about 10 per cent of the total sample from each year class for which plantings were made, except for the hatch of 1951. Samples from that year class have contained 14 per cent. The fall of 1950 was the only one during the period when adult trout were noted on the spawning beds. It has been estimated (Fry 1953) that egg deposition in that fall was of the order of 50,000 to 120,000 eggs. The increased incidence of unmarked fish from the 1951 hatch, although not statistically significant, is possibly, therefore, to be ascribed to this small spawning. It would appear then that it can safely be presumed that the majority of the unmarked lake trout that have been taken in South Bay in late years are planted fish that were improperly clipped or for which the clips were not recognized when they were handled after recapture. A summary of the numbers of lake trout taken in South Bay classified by marking and year class is given in Table II. Growth of Planted Fish

The age length relation of the various samples of the planted fish are shown in Figure I in comparison with the curve obtained for the native fish taken from Fry (1953). The ages were estimated from impressions on cellulose acetate read on various projectors or in some cases with a binocular microscope. The first year's growth showed the complex pattern recently described by Cable (1956). Scales from samples of the fish planted and of certain fish which had been tagged before planting and later recaptured, were available to enable the position of the first annulus to be placed. Little difficulty was experienced in reading the scales and internal evidence indicates that the readings are in general reliable. For instance no fish of the 1949 year class were planted and only four marked fish of the adjacent year classes were incorrectly 15 assigned to that group. In addition, as Figure I shows, the growth curve of a given year class tended to remain consistent in its course in comparison with other year

Figure I.--Grourth of South gay lake trout showing a comparison between the various year classes of hatchery stocked fish and the native stock. 16

classes. Note for example the curve for the 1948 year class. It can be seen that the growth of the planted fish has been in general the same as that which was displayed by the native stock. The 1948 year class (planted in 1949), which consisted of fish known to have been small when planted, although the sample taken for measurement was lost, showed the smallest size at a given age throughout its history. In addition to the growth curves in Figure 1, which are assessments based on scale reading, a direct comparison, based on marking and recovery data of the growth of planted fish for the year 1956, can be made with native fish for 1949. Twenty-seven of the fish released with tags in June of 1956 were recaptured in September and October. These fish had a mean length of 18.6 inches when released at a mean date of June 16. They were recaptured at a mean date of September 13 and had gained a mean increment in length of 1.3 inches. From Fry's (1953) Figure 2 it may be estimated that over the same period in 1949, native lake trout of similar size gained 1.1 inches. The samples of 1956 also gave an estimate of the intraseasonal growth of the age II individuals. Two samples of approximately 20 fish each were taken, one in June and the other in September. The mean length of the June sample was 9.5 inches, that of the September sample 11.5 inches.

Survival

In 1956 it was possible to make direct estimates of the sizes of the population of two of the year classes of planted lake trout, those of 1951 and 1952. These estimates were of the Petersen type and were made by releasing tagged trout taken in pound nets in the spring and recapturing a fraction in gill nets in late summer. The number of tagged fish released was 98 estimated to be age IV and 52 of age V. Of these, 15 age IV fish were recaptured with 276 untagged fish of the same age. The number of tagged fish of age V retaken was 10 together with 93 untagged ones. These data give Petersen estimates, in round numbers, of 2,000 age IV lake trout and 500 age V present at the time of tagging. These two estimates yield widely different percentage survivals from the two plants. Only two per cent of the 23,100 yearlings of the 1951 year class are thus estimated to have survived to age V, whereas 45 per cent of the 1952 year class of 4,500 planted are estimated to have survived to age IV. It has not been possible to make any further direct estimates of the number of survivors from other plants by the mark recapture method. However, some estimates are given below based on catch per unit effort.

Pound net catches Five pound nets have been fished in essentially the same locations in South Bay since fishing started in 1947. In certain of the earlier years additional pound nets were set, but for comparative catch-effort studies only the five basic nets have been considered. The catches of the five nets in question for the month of June are listed in Table III. June has been taken as the best month for comparison since the nets almcst invariably can be expected to take trout throughout that month. In some years it was not possible to get all nets set and fishing much before the first of June so comparisons of May fishing would include very irregular sections of that month. Also in 17 some years the trout retreat to deep water early in July, therefore weather might be expected to greatly influence comparisons made for that month. In general, pound net fishing for lake trout extends from mid-May to mid-July with June being the most productive period. Even in June it is possible that water temperature in some years has an influence on the pound net catch of trout; therefore the catch per unit effort of these nets cannot be considered a precise estimate of the relative size of the lake trout population from year to year.

The possibility of using the catch per unit effort to give rough estimates of the survival of the planted lake trout in South Bay rests on the series of population estimates made for the native lake trout population which still existed there until 1951. Population estimates of both the Schnabel and the Petersen types were made for the lake trout in the years 1948 through 1952 (Fry 1953). Petersen estimates were made again in 1956 as is discussed above. These various estimates, together with the catch per unit effort calculated for the same years, form the basis of the estimates of population size made for the other three years in which pound net catches were of any significance. Thus from the first five rows of Table III it may be estimated that, over the ages represented, a catch of one lake trout per lift represents a population of 775 lake trout present in South Bay. The data shown show no trend in the catch per unit effort with age so that it is not worth while to calculate the means for the individual ages. The population estimates given in the lower four rows of Table III are derived from the mean relation of C.U.E. and catch given above. In Table IV the population estimates for the planted fish given in Table III are related to the year classes. There is undoubtedly a great deal of uncertainty to be attached to these estimates. For example the estimate for the size of the 1951 year class at age IV is obviously wrong since approximately 200 individuals of this year class were taken in the sampling at ages IV and V after this estimate was made (see Table II). The season of 1955 in which the low population estimate was obtained was distinguished by being the warmest on record during the course of our observations. For that reason the pound net fishing for trout was terminated early, as the fish with- drew to deeper water, and the estimates based on the whole of the June fishery, would be expected to be low. In any event, while the percentage error may be gross, the values appear to indicate the order of magnitude of the various year classes in the fishery.

Percentage survivals to age V based on the estimates are shown in the last column of Table IV. A notable feature of the estimates is that the numbers surviving have borne no direct relationship to the numbers planted. Indeed, as far as the data go, the relation has been inverse which seems to indicate that at present the conditions in South Bay are such that no more than 1,500 or so lake trout can survive to age V regardless of the sise of the plant. It will be of interest to see whether such continues to be the case as the histories of the plants are followed further. This appears to be a rather surprising circumstance since earlier conditions permitted age V populations several times this number. The only obvious change in the environment which might be expected to affect the lake trout at sizes beyond those at which they were planted is the sea lamprey. 18

Another notable feature of Table IV is that fish of the 1948 year class did not survive to age VII, the age at which most females reach maturity, nor did the 1950 year class survive to age VI. These losses can both certainly be attributed to the depredations of the sea lamprey. This points out rather clearly why it is possible for the lamprey to completely eliminate a lake trout population by effectively halting all reproduction.

Lamprey Scarring

The scarring data are summarized in Table V which continues the type of summary given for the years 1947 through 1951 in Fry (1953) except that the size classes above 19 inches have been further condensed. The table is based on records of all fish handled during the period either killed or released. Fresh and healed scars have been combined.

Lamprey scarring on fish 19 inches and over has continued at the level attained in 1950 and was much increased in four of the later years. The apparent drop from the 1950 value in scarring on fish under 19 inches is due largely to the capture of many more fish under 16 inches in the net sampling during the years reported here. The 1950 sample was taken by angling and very few fish were captured that were under 18 inches. Thus many more fish of the lengths that are immune to scarring are included in the samples since 1950. In general the evidence from the scarring data indicates that the first year of substantial lamprey attack on the lake trout in South Bay was 1949 and that severe attacks began in 1950. The data for 1951 and 1952 are weak. Scarring probably continued at the 1950 level until the end of 1954. The severity of lamprey attack has further increased in the two last years of the records, 1955 and 1956.

The argument that the lack of healed scars on small lake trout is perhaps only an indication that virtually every attack at this size is fatal and that the mortality from this cause may be actually greater among small fish than among large ones, is not supported by the evidence gathered in South Bay. The South Bay data suggest rather that a wave of mortality (not due to lamprey) may come among the small fish early after planting and that the lake trout are then relatively immune to further losses until they approach twenty inches in length. In 1955, 89 fish from six to nine inches in length were tagged at the time of planting. During the 1956 fishing operations, nine of these tagged fish were recovered, ranging in size from 11 to 14 inches. The recovery of these tagged fish does not indicate any substantial mortality from lamprey attack among lake trout up to 14 inches long. It is rare that a lamprey scar is observed on fish of these lengths.

If the size composition of a year class of lake trout is followed through, year by year, in detail in South Bay, it seems that the loss by lamprey attack sharply increases as the fish approach a length of 20 inches, regardless of age. It is this circumstance apparently that has prevented the lake trout from living to be much older than age V in late years in South Bay. The single year class that reached age VI in substantial numbers, the 1948 year class, had apparently been retarded in the first and second year and did not attain the critical size as early as did the other year classes. 19 Conclusions

1. To date the plantings have shown that it promises to be quite feasible to rehabilitate a lake trout population in the Great Lakes by the planting of yearling stock. In the absence of lampreys two to 50 per cent of such planted fish could have been expected to have reached maturity in South Bay. It is possible, but not yet proven, that the maximum efficiency could be obtained with relatively small introduc- tions, for the data suggest that about the same absolute number of lake trout have survived regardless of the number planted. 2. The stocks introduced into South Bay have behaved in growth and distribution in essentially the same manner as did the native stock. 3. When sea lamprey are present in substantial numbers, planted lake trout can still survive in numbers for several years and apparently become highly susceptible to lamprey attack only after reaching a certain critical size which in South Bay is approximately 20 inches fork length. 4. It is improbable that any planted lake trout will survive to spawning in South Bay while the sea lamprey maintains its present population desntity there.

References

BLACK, E. C., 1956. Appearance of lactic acid in the blood of kamloops and lake trout following live transportation. Canadian Fish. Cult. 18:20-27. CABLE, LOUELLA E., 1956. Validity of age determinations from scales, and growth of marked Lake Michigan trout. U.S. Fish and Wildlife Serv. Fishery Bull. 107 (vol. 57) 55 pp. FRY, F. E. J., 1953. The 1944 year class of lake trout in South Bay, Lake Huron. Trans. Amer. Fish. Soc. 82:178-192.

Table I

Summary of Lake Trout Planting in South Bay, Lake Huron.

Year Class Number Planted Date Planted Average Length' Origin of Eggs Rearing Station (10071 'e) 1948 Oct. 1949 n(sm)d Lake Superior North Bay

1949 0 ......

1950 101 Oct. 1951 5.3 ( 68) Lake Superior Chatsworth

1951 231 Nov. 1952 3.7 (176) Lake Superior Sault Ste. Marie

1952 45 Oct. 1953 4.1 ( 41) Lake Simcoe and Chatsworth Lake Superior

1953 184 May 1954 3.6 ( 92) Lake Superior Chatsworth

1954 157 Nov. 1955 6.2 (185) Lake Superior Chatsworth

1955 510 (est.) May 1955 1.5 Lake Superior Sault Ste. Marie advanced fry, (no actual air dropped measurements)

1 Baeed on sample measured, number of fish in sample shown in parentheses. 59417-6-4 20

Table II

Numbers of Lake Trout Taken in South Bay in the Years 1952 to 1956. (The numbers in brackets are unmarked fish.)

Year Number Per cent Class Planted II III IV V VI VII Total Unmarked (100's)

1947 0 . 1(0) 1(1) 7 1948 71 ...... 0(14) 79(2) ...... 2'74 11 1949 0 0(4) 2(2) ...... 10 1950 101 2(0) 12(3) 85(8) 20(2) 9(1) ...... 142 10 1951 231 0(1) 24(6) 25(4) 49(22) . 14 1952 45 17(5) 63(14) 86(40) ...... 11 1953 184 30(5) 193(18) ...... 9

1954 157 140(13) ...... 153 9

Total ...... 1588 ......

Table III

June Catches of Lake Trout in South Bay Pound Nets.

June Catch/life C.U.E. by Age Gronps 'Population Estimates Year Catch Lifts orC.U.E. IV V VI VII IV V VI VII

1948 44 1383 31.4 27.3 4.0 0.1 ...... 16,700 2,500 300 1949 132 1559 11.8 2.7 9.1 ...... 3,200 10,800 ...... 1950 69 477 6.9 ...... 1.7 5.0 0.2 ...... 1,000 3,000 160 1951 84 50 0.6 ...... 0.1 0.5 ...... 125 475 1956 50 160 3.2 2.1 1.1 ...... 2,000 500 ......

Based on C.U.E. and above estimates 1952 Insufficient data

1953 39 56 1.4 ...... 1.4 ...... 1 400 ...... 1954 30 40 1.3 0.4 ...... 0.9 ...... 310 ... 1955 51 17 0.3 0 .1 0.2 ...... 80 1 ......

1 Estimates based on mark recapture data prior to 1956 from (Fry 1953). 21 Table IV

Population Estimates in Relation to Age for Various Year Classes of Planted Fish.

Age Year Class Per Cent Survival IV V VI VII to Age V

1948 ...... 1 100 700 0 15 1950 ...... 310 160 0 2 1951 ...... 80 500 2 1952 ...... 2,000

Table V

Percentages of Lake Trout from South Bay Examined Bearing Lamprey Marks, a dash indicates a negative finding on a sample of less than 10 fish.

Fork Length, Inches Year Under 19 19-21 22 and Over

1949 ...... 1 11

1950 ...... 14 25

1951 ...... 37

1952 ...... 19 1953 ......

1954 ......

1955 ...... -- 71(7 fish)

1956 ...... 85 (13 fish)

59417-6-4+} T Notes on the Food of the Young of Three Species of Pacific Salmon in the Sea by

Murvel E. Annan University of Washington

During July, 1950, very considerable numbers of young pink, chum and spring Pacific salmon were taken in beach seining operations in the San Juan Island area, carried out as part of the work in a course in Fish Biology given at the Oceanographic Laboratories of the University of Washington at Friday Harbor, Wash. Cursory examination of some stomach contents of these fish revealed the presence of dipterous larvae and pupae in addition to crustaceans, and so from July 10 to 31, 206 specimens were preserved for more detailed examination.

In the final analyses of the data, no differences were found in the nature of the foods among the three species of salmon or in relation to the size of the fish. The data were therefore not segregated.

The species were represented as follows: pink salmon, Oncorhynchus gorbuscha 103; chum salmon, O. keta 54; and spring salmon, O. tshawytscha 49; total 206. The total lengths of the fish ranged from 4.0 to 9.5 cm., with a mean length of 6.5 cm.

The food materials of the total number of fish were estimated to be: Copepoda 50 per cent; Diptera (larvae and pupae) 26 per cent; miscellaneous items including small crustacea (copepods, isopods, amphipods) and unidentifiable material 24 per cent.

Since the presence of Diptera was surprising, special attention was given to the identification and occurrence of this food item.

Reference to publications by Saunders (1928) and Johannsen (1936) readily led to the identification of the insect as Camptocladius pacificus Saunders in the family Chironomidae. However, systematists at the present time appear to prefer to consider Camptocladius as a sub-genus of the genus Spanistoma (from correspondence with the U.S. National Museum). On this basis the name is Spanistoma pacifica (Saunders).

Examination of larvae and pupae was made after clearing specimens in gylol and mounting in paramount. Drawings are presented in Figure 1.

The data on occurrence are given in Tables I and II. When the Diptera were more abundant than other food items, they were recorded as predominant. Fewer Diptera were eaten as the season progressed. Either they became fewer or other food forms became more plentiful. The writer is indebted to Dr. W. A. Clemens for suggesting the study.

23 24

Figure 1.-Spanistoma pacificus (Saunders). I Pupa. II Anal pseudopod showing hoolls. III Lanra. IV Mandible. V Labium. Drawings made with the aid of a camera lucida. Table I Diptera, Spanistoma pacifica, in the Stomachs of Young Salmon.

Number Number Number Having Per Having Per Date of Salmon Diptera Cent No Cent Predominant Diptera

10 Ju1y ...... 7 4 57 0 0 10 July ...... 32 18 56 3 9 10 July ...... 20 14 70 2 10 10 July ...... 17 4 24 3 18 10 July ...... 8 2 25 1 13 17 July ...... 18 0 0 10 56 17 July ...... 35 11 31 6 17 17 July ...... 8 1 13 4 50 24 July ...... 6 0 0 4 67 24 July ...... 11 0 0 7 64 24 Ju1y ...... 1 0 0 1 100 31 July ...... 28 0 0 24 86 31 July ...... 15 0 0 15 100 206 54 26 80 39 25 Table II

Occurrence of Diptera, Spanistoma pacifica, in the Stomachs of Young Salmon by Weeks.

Number Number Number Having Week of Per Having Per of Salmon Diptera Cent No Cent Predominant Diptera

10 July ...... 84 42 50 9 11

17 July ...... 61 12 20 20 33

24July ...... 18 0 0 12 67

31 July ...... :...... 43 0 0 39 91

206 54 26 80 39

Literature Cited

JOHANNSEN, O. A. 1936. Aquatic Diptera Part III, Chironomidae. Cornell University Agricultural Experiment Station.

SAUNDERS, L. G. 1928. Some Marine Insects of the Pacific Coast of Canada. Annals of the Ent. Soc. of America. 21(4): 521-545. I I A Direct-Current Electrofishing Apparatus Using Separate Excitation by

A. R. Murray

Fisheries Research Board of Canada, Biological Station, St. John's, Newfoundland

In order to determine the standing crops of Atlantic salmon, brook trout, and eels in the Little Codroy River, Newfoundland, a direct-current electrofishing apparatus using separate excitation has recently been devised. Separate excitation is a simple and practical method for getting excellent output voltage control. It is also a safer and more economical means of producing power for electrofishing. The purpose of this paper is primarily to describe the unit and report briefly on some results of its use.

The electrofishing apparatus consisting of engine, generators, control panel, cable and reel, positive fishing electrode, and negative ground electrode, is shown in Figure 1A. The source of power is a 6.5 h.p. Briggs and Stratton air-cooled, gasoline engine. An engine of this size was selected in order to take advantage of its relatively low r.p.m. characteristics (2,400 r.p.m.). The generators are belt driven (Figure 1B), with the exciter generator turning at 1,440 r.p.m., and the main generator at 1,250 r.p.m.

The wiring diagram for the electrofishing apparatus is shown in Figure 2. The main generator (Figure 1C) is a rebuilt, separately excited, direct-current unit mounted in an "Imperial Electric" frame. The armature is capable of producing a maximum unloaded potential of 675 volts at 3 amperes continuous service, or 4 amperes inter- mittent service. The resistance of the armature is 4 ohms. The field is wound for separate 12-volt excitation with an external field rheostat mounted in the control panel for output control. The resistance of the field is 36 ohms. Interpole windings are I used to improve voltage regulation under varying load conditions. The exciter generator (Figure 1D) is a Lucas 12-volt, shunt-wound, direct-current First, excellent I generator. There are three reasons for using separate excitation. voltage control can be obtained using separate excitation (along with interpole windings) because field troubles from variable voltage and load are minimized. Separate excitation means that the main generator has a low voltage separately excited field (i.e., 12 volts, as compared with 675 volts if the generator were self-excited). This minimizes the danger of insulation breakdown and flashover in the damp conditions existing beside rivers. Second, safety is a feature of separate excitation because 12 volts are used to control 675 volts. There are actually two safety factors involved: (a) the 12 volts separate excitation, and (b) the relay which is mounted in the control panel. The 12-volt generator is the source of power for the relay. The switch in the handle of the positive electrode (Figure 2) controls the relay, and the relay, in turn, controls the 27 28

field of the main generator. Thus there is just enough low voltage current flowing through the electrode switch to operate the relay. Third, because the resistors, relay, and switches are in the 12-volt circuit they are low priced and easily obtainable. The control panel (Figure lE) consisting of field rheostat, voltmeter, ammeter, 12-volt relay, and main switch, is mounted in a standard steel junction box. The 29

POSITIVE ELECTRODE

Field Rheostat

Main Switch

3-Conduetor Cabtyre

Armature (4A)

Single Conductor Cobty►e EXCITER = GENERATOR MAIN GENERATOR }

NEGATIVE ELECTRODE

Figure 2.-Schematic wiring diagram of direct-current electrofishing apparatus. 30

control switch on the 13-ohm field rheostat allows the selection of any one of 5 output voltages (unloaded values) : 200, 300, 400, 520, and 675. The variable output voltage is a useful feature for it allows the selection of the voltage best suited to the con- ductivity of the water, as well as the size and species of fish being collected. The control panel is separate from the engine and generators, the circuits being completed by means of two short cables from the control panel plugged into polarized receptacles on each generator. The engine and generators are bolted to a light weight tubular steel (bicycle-type) frame made in two sections, one carrying the engine and exciter generator, and the other the main generator. The weight of the engine and exciter section is 168 pounds, and the weight of the main generator section is 144 pounds. The two sections are carried separately (Figures 1C and 1D), but are bolted together when the unit is in operation (Figures 1A and 1B). The main generator can be moved back and forth on its frame to allow for belt adjustment. As suggested by Smith and Elson (1950), the engine and generators are mounted on two frames for convenience in handling because of their weight. However, an overall improvement in portability has not resulted because of their bulk and also because of the weight of some of the accessory gear (such as the barrier nets). Hence it is possible to use this equipment only in the areas of the river which are accessible by vehicle. The cable reel (Figure 1A) is a metal flanged drum which carries 300 feet of 3-conductor, number 14, cabtyre cable. One end of the cable is plugged into a polarized receptacle on the control panel, and the other end into a polarized receptacle on the handle of the positive fishing electrode. In practice, the reel is mounted on a horizontal axle, which is supported by a stand made of steel pipe, and the approximate amount of cable required is pulled from the reel and stretched along the stream bank. The man operating the fishing electrode drags the cable behind him. The negative pole of the main generator is grounded by connecting it through 100 feet of single-conductor, number 14, cabtyre to a 3- by 10-foot piece of copper hardware cloth which is set on the stream bottom and covered with rock or gravel. The single-conductor cable is also carried on the reel. The total weight of the reel and both cables is 114 pounds. The positive fishing electrode (Figure 1F) consists of a straight piece of 8 inch copper tubing, on the distal end of which is a rectangular piece of 2 inch copper tubing, size 8 by 11 inches. The proximal end of the electrode is attached to a tubular insulated handle by means of a brass coupling. There is a sliding contact switch on the handle which indirectly controls the main generator through the relay, hence the high potential will not be produced unless both the electrode switch and relay are closed. All members of the electrofishing crew wear electrician's rubber gloves and rubber chest waders. A Jeep and 2 wheel, â-ton trailer are used to transport the crew, apparatus, and accessory equipment. The unit has been used under a wide variety of conditions on the Little Codroy River, and has proven highly satisfactory. It has, for example, been used in both winter and summer in water ranging from 2.8°C. to 19.2°C. It has been used in various parts of the drainage basin in water the resistivity of which has varied from 19,800 31 to 42,200 ohms per centimeter cube (St/cm.8). The positive fishing electrode has been fished up to 400 feet from the negative ground electrode with no apparent decrease in efficiency. The total numbers of fish captured at the various sampling stations are shown in Table I. Because this work is in its preliminary stages, a more complete analysis is not given. The data are included only to indicate that the apparatus has been successfully used to catch Atlantic salmon parr, brook trout, and eels under a variety of conditions. The method of blocking off the sample areas with barrier nets, and the fishing procedure used at each station were much the same as that described by Godfrey (1956). Observations on the behaviour of Atlantic salmon parr, brook trout, and eels in the direct-current field parallel those of Godfrey. As reported by Saunders and Smith (1954), it was found that collection of fish was easiest when the generator voltage was adjusted so that the fish could be held at the positive fishing electrode in a condition of partial paralysis. The potential was around 300-375 volts. In general, voltages around 300 were found to be more satisfactory in warm water, whereas voltages around 375 were more satisfactory in cold water. This complements the findings of Smith and Elson. Table I

Numbers of fish taken at each of the eampling areas, Little Codroy River, 1956.

Atlantic Salmon Parr

Station Number Yearlings Brook Trout Eels Under- and yearlings Older

2 ...... 38 0

3 ...... 26 1

5 ...... 31 1

6 ...... 63

8 ...... 237

9 ...... 1

16 ...... 92

16a ...... 110

17 ...... 0

18 ...... 56

19 ...... 29

In connection with voltage adjustment, it was observed that the lower the resistivity of the water, the lower the voltage produced, and the greater the current drawn, at any one setting of the control selection switch on the field rheostat. For example, when the control selection switch was set so that the main generator was 32 producing 650-675 volts (no load), under load in water of 19,800 SZ/cm.$ resistivity it was producing 330 volts at 1.7 amperes, and in water of 42,200 SZ/cm.3 resistivity it was producing 600 volts at 0.9 amperes. The effect of water temperature and water resistivity on voltage adjustment were also observed in a general way. At any given resistivity, the higher the water temperature, the lower the voltage produced, and the higher the current drawn. For example, with the control selection switch set as above, the main generator produced 500 volts at 1.2 amperes in water of 40,200 12/cm.3 and 15.3°C., whereas it produced 600 volts at 0.9 amperes in water of 42,200 Sl/cm.3 and 4.4.'C. Some immediate mortality was observed in two instances when fishing in rapid water over rubble bottom when dislodged and shocked parr drifted downstream into the lower barrier net and were held by the pressure of the water and by the accumulation of debris on the net. This occurred early in the season when the apparatus and its attendant techniques were new to the crew. As the season progressed and the crew became more proficient, there was no further immediate mortality. Shortly after the unit was completed in January 1956, it was tested in the Salmonier River, St. Mary's Bay, Newfoundland. Approximately 160 Atlantic salmon and brown trout parr were collected. These fish were brought back to the Biological Station and held for observation in wooden troughs and glass aquaria for a period of one month. There were no mortalities during this period. Accordingly, delayed mortality has been assumed to be not significant, and it has further been assumed that fish which recovered immediately from the effects of the electroshock were uninjured. Similar conclusions have been drawn by Smith and Elson.

Acknowledgements

Assistance in the design of the apparatus by Mr. W. Brown, of the Department of Transport, St. Andrews, Newfoundland, Mr. R. C. Hawksford, of the Vocational Institute, St. John's, and Mr. M. A. Foley, of the Fisheries Research Board of Canada, Technological Unit, St. John's, is gratefully acknowledged. The apparatus was built by Canadian Machinery and Industry Construction Limited, St. John's. This paper is published with the permission of the Fisheries Research Board of Canada.

Literature Cited

GODFREY, H., 1956. Catches of fish in New Brunswick streams by direct-current electrofishing. Can. Fish Cult. 19: 1,8.

SAUNDERS, J. W., and M. W. SMITH, 1954. The effective use of a direct-current fish shocker in a Prince Edward Island stream. Can. Fish Cult. 16: 4249.

SMITH, G. F. M., and P. F. ELSON, 1950. A direct-current electrical fishing apparatus. Can. Fish Cult. 9: 34-46. Back-pack Fish Shocker by

A. A. Blair

Fisheries Research Board of Canada, Biological Station, St. John's, Newfoundland.

The value of small, portable, electrical fish shockers is fast becoming apparent to fishery workers. The immediate need for such a unit in Newfoundland was for collecting live salmon parr for laboratory experiments on tagging. The intention was to assemble a unit similar to one described by Haskell et at. (1954). Difficulty was experienced in obtaining the specified parts locally, but this was overcome by obtaining a packaged D.C. unit, the Vibrapack, with a similar circuit from P. R. Mallory & Co. Inc., Indianapolis, Indiana.

Description

The complete shocker is shown in Figure 1. It consists of a 6-volt motorcycle battery, the Vibrapack, and a voltmeter housed in a fairly water-tight box and two paddle-type electrodes. The outside dimensions of the box are 15 inches wide, 15 inches high, and 8; inches deep. A pack-sack frame is attached to the box so that the outfit can be carried comfortably on the back. The electrodes, which are made of â inch copper tubing, are 42 feet long, the round end being 7 inches in diameter. The grid in the round end is formed of heavy copper wire. A switch in the low voltage D.C. circuit for starting and stopping the unit is conveniently located on the handle of the right or positive electrode. A switch for use as a main switch or in emergencies is situated on top of the case. It has a pull cord which is readily accessible to the operator or the second man of the team. The unit SHOULD NEVER be used by one man alone. The total weight of the outfit is 44 pounds, so it can be carried fairly easily by one man. The cost in St. John's, Newfoundland, was $120.00. (It should be less in most other cities in Canada.) The following is a list of parts and materials used:-

Parts Battery ...... 6-volt motorcycle battery, Lucas PUZ 7E-9 Vibrapack ...... Mallory VP-555H Voltmeter ...... 0-500 V.D.C., 1,000 ohms per volt, Triplett Main fuse ...... 30 amp. Vibrapack fuses...... 10 amp. 3 AG Switch No. 1 ...... Open knife, SPDT, mounted on outside of case

33 34

Switch No. 2 ...... Slide switch, SPST, mounted on positive electrode handle Relay ...... Automobile horn relay, Sorensen HR-2 Discharge resistor R-4...... 400 K safety resistor

Materials

Electrodes ...... zinch copper tubing Electrode handle insulation. .... g inch automobile heater hose Electrode connections ...... For positive electrode, three conductor No. 12 cab tyre For negative electrode, single conductor No. 12 cab tyre Battery connections ...... No. 6 insulated flexible cable

Figure 1.-On the left a view of the shocker ready for use, and on the right a view of the interior. 35 BATTERY

S-I

RELAY

VIBRATORS

TRS.

• ------^ ^------

r------^ ^------,

S-2

...... VIBRAPACK...... : I ELECTRODES

Figure 2.-Schematic wiring diagram. Symbols: B-, battery negative; B+, battery positive; C-1, C-2, C-3, C-4, condensers; CH-2, CH-3, chokes; FIL., filament; G, ground; R-2, R-3, R-4, resistors; S-1, S-2, switches; TRS., transformers: VIB., Vibra pack; and Q voltmeter. 36

The Vibrapack or main power unit is factory built. Types are available for nominal input of 6, 12, and 32 volts, nominal output ranging from 125 to 400 volts, and maximum output of 60, 100, 150, and 200 milliamperes. The one selected for operation in Newfoundland streams has a nominal input of 6 volts, a nominal output of 300 volts, and a maximum output of 200 milliamperes. The wiring diagram of the Vibrapack unit is shown within dotted lines in the schematic wiring diagram of the shocker (Fig. 2). For ease in assembly and to give good connections all wires have soldered terminals. The Vibrapack is supplied with OZ4A rectifier tubes which according to the manufacturer's description require a minimum current load of 140 milliamperes to start the unit. However, in this particular installation it was found that the unit would start with a load as low as 90 milliamperes.

39C

U ci 360 S V ^-N 0 33 ss M^k/ti,^ > k M

D 300 a SQ I -- D 0 270

240

20 5`V 5.4 S.SV. 5^5

.-,_MINIMUM OUTPUT CURRENT 40 MA. z 2 75 100 125 150 175 200 OUTPUT MILLIAMPERES

Figure 3.-Operating characteristics of the shocker in 6 inches of water with distances of 6 inches to 4 feet between eleccrodes.

Operation The power to operate the shocker is supplied by a 6-volt motorcycle battery. This direct current is changed successively as follows: to low voltage alternating current by a vibrator, to high voltage alternating current by a transformer, and to high voltage direct current by a rectifier. This high voltage direct current is conducted to the electrodes for shocking the fish. The operating characteristics of the shocker 37

(Fig. 3) were determined in 6 inches of water in a wooden trough, 4 feet long and 14 inches wide, with distances of 6 inches to 4 feet between electrodes. All readings depend on electrode separation and immersion. The drain on the battery ranged from 5.1 to 5.75 volts and from 6 to 20 amperes. The output of the unit ranged from 240 to 415 volts and from 40 to 200 milliamperes. When the battery cannot produce 18 amperes at 5.1 volts, the unit will not start. If the output drops below 40 milliamperes at 240 volts, the unit will cease to operate. If the electrodes should accidentally come closer together than 6 inches at any time, a blown fuse will result, safeguarding the unit from overloading and damage to the component parts. The shocker has worked quite well in the streams fished so far where depths ranged from 6 inches to 3 feet. Fishing downstream was found to be more effective than fishing upstream. A team of two men is required, one man to operate the shocker and the second man to do the collecting. In clear water with little current, the collector used a dip net and moved along with the operator. In turbid or swift water, the collector used a one-man seine and took up a stationary position while the operator started about 20 feet above and worked down to the seine. Fish were narcotized within 1 foot of the positive electrode. Beyond that, some were attracted to the positive electrode but most appeared to receive a slight shock and moved in all directions. Occasionally fish were led into the seine. All fish recovered quickly. Since the unit has been quite effective in occasional collections, it shows that a current of 40 to 200 milliamperes is ample for small stream sampling in soft waters. The unit has not been used in hard waters, which would require a higher output current (Haskell et al., 1954). Two fully charged batteries have supplied enough power for a full day's fishing. One battery will operate the unit continuously for 1 hour. Under actual fishing conditions where the unit is not in continuous operation, it is estimated that the battery will operate the unit effectively for a maximum of 2 hours. Fishing time is thus limited by the number of batteries available or distance travelled from a charger. A word of caution is necessary in connection with shutting off the unit. This should be done while the electrodes are immersed in the water so as to drain of any current left in the condensers. Otherwise the resultant shock upon contact with any of the uninsulated parts is quite uncomfortable and may cause other accidents. This danger could be eliminated by inserting a discharge resistor of 400 K across the output leads. It is shown as R-4 in the schematic wiring diagram (Fig. 2).

Acknowledgements

The fish shocker described in this paper was assembled by Mr. A. P. Cowan who is a technician on the staff of the Fisheries Research Board of Canada Biological Station at St. John's, Newfoundland. He also prepared the figures. Mr. W. M. Gaskell, Electrical Equipment Technician with the Newfoundland Light and Power Company, Limited, very kindly checked the technical parts of the paper.

Literature Cited

HASKELL, DAVID C., JOHN MAcDOUGAL, and RODERIC FREEMAN. 1954. Two back-p4c}1 fish shockers. New York Fish and Game J., 1(1): 65-74.

Experiments with Toxaphene As Fish Poison by

George E. Stringer and Robert G. McMynn

British Columbia Department of Recreation and Conservation, Fish and Game Branch, 567 Burrard Street, Vancouver, British Columbia.

Introduction

Elimination of undesirable fish species and subsequent re-introduction of preferred species is an effective means of improving sport fishing. Since this form of management is expensive, the volume and number of lakes which can be treated is limited. If a cheaper chemical treatment could be proven, lake rehabilitation through poisoning would become more popular. In this report, the commercial insecticide, Toxaphene (a chlorinated camphene) has shown some promise when tried in a series of alkaline lakes in British Columbia. Although some information is available concerning the chemistry of Toxaphene and its effects on warm-blooded animals, little critical work has been done on its use as a fish toxicant. Since British Columbia contains many lakes in which the sport fishery would benefit by removal of predator and competitor fish, it seemed desirable to determine the effectiveness and limitations of Toxaphene both with respect to fish as well as to plankton and bottom fauna. To obtain this information, eight small alkaline lakes typical of those found in the interior of the province were chosen. They contained a variety of coarse fish, all were readily accessible, and with one exception, had no permanent inlet or outlet streams.

Methods

In 1954, several British Columbia lakes were treated with "Fish Tox" (a rotenone based chemical containing 10 per cent Toxaphene). These lakes remained toxic for over two years. The prolonged period of toxicity was attributed to the presence of Toxaphene. As a result of this observation, a series of four Toxaphene concentrations was chosen in this experiment ranging downward from 0.10 p.p.m. (concentration represents actual concentration of 100 per cent Toxaphene). The concentrations, 0.01 p.p.m., 0.03 p.p.m., 0.07 p.p.m. and 0.10 p.p.m. were duplicated. Some of the more important physical and chemical features of the experimental lakes are listed in Table I.

39 40

Table I

Physical and Chemical Features of Toxaphene Treated Lakes

Toxaphene Name of Concen- Area Mean Depth Volume T.D.S. Lake tration (acres) (feet) (acre feet) pH (p.p.m.) (P.p.m.)

Alleyne...... 0.01 135 55 7,425 8.3 308

Taylor...... 0.01 14 32 463 7.9 249

Gladstone...... 0.03 22 25 533 7.9 277 Round...... 0.03 91 58 5,278 8.2 295 Gallagher...... 0.07 17 34 571 8.2 172

Spectacle...... 0.07 9 11 96 8.4 285

Summit ...... 0.10 12 30 360 7.9 249 Lady King'...... 0.10 15 18 270 8.2 170

1 Lady King has an outlet stream discharging about two cubic feet per second.

In addition to the physical and chemical surveys of the lakes, bottom fauna and plankton samples were taken prior to treatment. Poisoning of the lakes took place during the first week of August, 1956. Gill nettings, live cage sets (rainbow trout), plankton hauls, and bottom samplings were carried out after the poisonings at an interval of one month and again at nine months. Periodic checks will continue until the lakes become non-toxic to fish. Initially, when mixing and handling containers of liquid Toxaphene, workers were completely enclosed in rubber clothing including goggles and respirators. Toxa- phene splashed on the skin produced no harmful effects provided that it was immediately washed off with a detergent. Respirators reduced the possibility of inhaling fumes while transferring liquid from one sealed contained to another. When applying the concentrate, only goggles, rubber boots and gloves were found necessary as additional protective clothing. Although all experimental lakes he within cattle raising country and are used constantly as water holes, no adverse effect on cattle has been noted. Reasonable caution makes the handling and distribution of Toxaphene no more hazardous than other similar chemicals in daily use. Liquid Toxaphene was easily and quickly distributed by the use of an outboard motorboat, fire pump, a 10 or 20 gallon drum and fittings. The required amount of Toxaphene for each lake was measured into 10 or 20 gallon drums and sealed before being transported to the lake. Each drum was provided with two outlets, one to a shut-off valve and the other to a capped breather pipe. Neoprene connections led from the shut-off valve, first to a needle valve which provided a fine flow adjustment, and then to a domestic water meter. This meter recorded volume delivered but not rate. The latter was calculated from timed runs using water and allowing the pump to operate at capacity. The metered line from the drum was then bled directly into 41

the suction line of the firepump, which drew water from the lake continuously and at capacity (50 to 75 gallons per minute). Discharge of the mixed dilute Toxaphene was effected by two lines from the pump which released the mixture underwater on either side of the boat. Toxaphene Effects Fish

In order to subject as many fish species as possible to the various Toxaphene concentrations, cages containing fish not occurring in the lake were used to augment indigenous species. These were:-

1. Largescale sucker (Catostomus macrocheilus) 2. Squawfish (Ptychocheilus oregonense) 3. Carp (Cyprinus carpio) 4. Redside shiner (Richardsonius balteatus) 5. Perch (Perca f iavescens) 6. Peamouth chub (Mylocheilus caurinum) 7. Lake chub (Couesius plumbeus) 8. Mountain Whitefish (Coregonus williamsoni) 9. Kokanee (Oncorhynhus nerka) 10. Rainbow Trout (Salmo gairdneri) 11. Prickly sculpin (Cottus asper)

At time of treatment, all lakes were thermally stratified with surface temperatures ranging between 64°F. and 67°F. In general fish displayed distress symptoms (erratic swimming, surfacing, and little alarm reaction) within four hours of treatment. Within 48 hours, the majority of fish were dead at all concentrations of Toxaphene; higher concentrations killed more quickly. At 0.10 p.p.m. all caged fish (suckers, squawfish, carp and peamouth chub) had succumbed prior to checking at 48 hours. All caged fish at a Toxaphene concentration of 0.10 p.p.m. died before 96 hours had elapsed. It was noted that a few individuals of a species exhibited remarkable ability to resist the lethal effects of Toxaphene for a prolonged period, while the remainder of the individuals of the same species died more quickly. In Gallagher Lake, one whitefish and one sucker were netted 19 days after the chemical treatment; in Spectacle Lake two caged carp were alive after four days (one was still alive one month later, but died before being checked at six weeks). In Alleyne Lake, one live sucker was netted three weeks after treatment. Gill netting, (; inch to 32 inch mesh) undertaken two to three weeks after treatment showed that all fish had been killed, with the exception of those in the above-mentioned lakes. Simultaneous live cage tests with rainbow trout in Alleyne and Taylor Lakes (0.01 p.p.m.) produced 100 per cent mortalities within a five day period, while controls in adjacent lakes produced almost 100 per cent survivals. 42

In May, eight to nine months after the poisonings, gill net sets did not produce a single fish in any of the treated lakes. Live cage tests (rainbow trout) produced 100 per cent mortalities within two to seven days, with one exception, in all lakes. Although some of the cages were not checked until the seventh day, it was obvious that the fish had been dead for some time. In Lady King Lake two live cages were utilized; one was set at a depth of five feet and the other at a depth of about 15 feet. After seven days the five foot cage contained no live fish, while the 15-foot cage had a 60 per cent survival after 13 days. This particular lake is the only experimental lake which has a continuous inflow and, since the temperature of the inlet stream (55°F. on June 5) was considerably cooler than that of the lake surface (64°F. on June 5), it is probable that, owing to density difference, the non-toxic inlet waters diluted or replaced the deeper chemically treated waters thereby creating a more favourable enviroment for the survival of fish at that depth. Tables 2 and 3 sum- marize the results of the Toxaphene treatments on fish.

Table II

Results of Toxaphene Treatments on Fish Within 4 to 120 Hours After Poisoning.

Name of Lake Fish Present Initial Results

Alleyne ...... Sucker 24-48 hours. Many dead suckers (0.01 p.p.m.) R. Trout some still in distress

Taylor ...... Sucker 1 Peamouth chub 1 96 hours. All fish in cages dead. (0.01 p.p.m.) Squawfish I R. Trout Carp 3 Shiner

Gladstone...... Sucker 1 Peamouth chub 1 12 hours. Dead shiners and trou (0.03 p.p.m.) Squawfish i R. Trout many fish in distress. Carp i Shiner

Round ...... Carp 72 hours. No live fish observed. (0.03 p.p.m.)

Gallagher ...... Sucker Peamouth chub 96 hours. No live fish observed. (0.07 p.p.m.) Squawfish Lake chub Shiner Whitefish Perch Kokanee R. Trout

Spectacle ...... Sucker 1 Peamouth chub 1 120 hours. All caged fish dead. (0.07 p.p.m.) Squawfish 1 Shiner Except 2 carp 4. Carp

Summit ...... Sucker Peamouth chub 4 hours. Many dead trout and (0.10 p.p.m.) Squawfish R. Trout shiners. Distress noted in Carp Shiner others. 48 hours all caged fish dead.

Lady King ...... Squawfish R. Trout 24 hours. No sign of fish life. (0.10 p.p.m.) Shiner Sculpin Peamouth chub

I Fish held in cages. 2 Dead in 1 to 2 months. 43 Table III

Results of Toxaphene Treatment on Fish 2 to 3 Weeks and 8 to 9 Months After Poisoning.

2-3 Weeks 8-9 Months After Treatment After Treatment Name of Lake Fish Present Gill Net Live Cage Gill Net Live Cage Sets Tests Sets Tests

Alleyne..... Sucker 21 days After five No fish Fifty days (0.01 p.p.m.) R. Trout one sucker2 days all dead 100 1^-C mort. Taylor...... Sucker' Peamouth chub' 20 days no After five No fish Seven days (0.01 p.p.m.) Squawfish1 R. Trout fish days all dead 100% mort. Carp' Shiner

Gladstone. .. Sucker' Peamouth chubt 20 days no No fish Five days (0.03 p.p.m.) uap wfish' fish 100 ô mort.

Round...... Carp 16 days no No fish Two days (0.03 p.p.m.) fish 100177C mort. Gallagher.. .. Sucker Peamouth chub 19 days, one No fish Seven days (0.07 p.p.m.) Squawfish Lake chub whitefish,2 100c- mort. Shiner Whitefish one sucker= Perch Kokanee R. Trout

Spectacle. ... Sucker' Peamouth chub' 19 days no No fish Seven days (0.07 p.p.m.) Squawfish' Shiner fish 100 C, mort. Carp'

Summit. .... Sucker' Peamouth chub' 20 days no No fish Two days (0.10 p.p.m.) Squawfish' R. Trout fish 100c- mort. Carp' Shiner

Lady King ... Squawfish R. Trout No fish Seven days (0.10 p.p.m.) Shiner Sculpin 1001 mort. Peamouth chub and 40% mort.

' Fish held in cages. = Re-netted 74 days after treatment-no fish

Bottom Fauna

An evaluation of the effects of Toxaphene on bottom organisms was attempted. Samples (six inch Ekman dredge) consisting of 30 pre-poisoning and 70 post-poisoning dredgings were taken from each lake. Limitations of this type of sampling preclude quantitative analysis of the data. Qualitative results, however, are summarized in Table 4.

Shrimp (Gammarus or Hyallela.) were taken from six of the eight lakes prior to treatment, but were absent in subsequent dredgings following poisoning. Further observations on the sensitivity of shrimp to Toxaphene were provided by placing shrimp from a nearby lake into cages in Round and Lady King Lakes eight to nine months after poisoning. The cages were set at a depth of about 15 feet and checked after a period of six days. All shrimp in the Round Lake cages were dead; no mor- 44

talities were noted in the cage in Lady King Lake. The results from Lady King are undoubtedly associated with the inflowing stream and the depth at which the cages were set Table IV

Occurrence of Bottom Organisms in Toxaphene Treated Lakes Before and After Poisoning.

('Y' Indicates Presence of the Particular Form "-" Indicates that the Lake Was Not Sampled, While "o" Indicates that the Organism Was Absent in the Samples).

Alleyne Taylor Gladstone Round Gallagher Spectacle Summit Lady King Organism (0.01) (0.01) (0.03) (0.03) (0.07) (0.07) (0.10) (0.10) ABC ABC ABC ABC ABC ABC ABC ABC

Shrimp...... - oo x oo - 00 x - o x oo xoo xoo xoo

Midge Larvae .... X x x x x x x x x x- x x x x x x x x x o x x o Aquatic Earthworms.... xxx xxx xxx x - x x xx xxx xxx xxx

Leeches...... o00 00o xx o o - o 0 0o xox xxx oox

Mayfly Nymphs.. o 0 o x x o x 00 x- o x o o x x x x o 0 00 0 Dragonfly and Damselfly Nymphs...... xoo xxx xxx x - o x oo xxx xoo 000

FreshwaterSnails. xxx xxx xxx x - o x xx xxx xxo xxx

A-Before treatment. B-One month after treatment. C-Nine months after treatment.

Midge larvae (Chironomidae) were present in all lakes initially as well as one month after application of Toxaphene. Nine months after treatment midge larvae were absent from Lady King and Summit Lakes (0.10) but present in the other lakes. This indicates that at a concentration of 0.10 p.p.m. Toxaphene, midge larvae are either eliminated or greatly reduced in abundance, with death apparently occuring somewhere between one and nine months. Aquatic earthworms (Oligochaeta) were present in all lakes before and after treatment. Leeches (Hirundinea) although not taken in all samples were found in the two 0.10 p.p.m. Toxaphene treated lakes nine months after poisoning. It would therefore appear that these forms are unaffected at even the highest concentration of Toxaphene used in the experiments. Mayfly nymphs (Ephemeroptera) were taken in six of the lakes prior to treatment but were absent in subsequent samples from all lakes except Taylor Lake (0.01 p.p.m.). This suggests that mayfly nymphs are killed by Toxaphene concentrations of 0.03 and greater. Dragonfly and damselfly nymphs (Odonata), although taken initially in most lakes, were absent in later samples taken from both the 0.10 p.p.m. and one of the 0.07 p.p.m. 45

treated lakes. It would seem that a concentration of 0.07 p.p.m. is near the upper limit of tolerance for this group. Lethal effect of 0.10 p.p.m. concentration was amply demonstrated in the case of Summit Lake where numerous dead dragonfly larvae were found one month after application of 0.10 p.p.m. Toxaphene. Freshwater snails (Gastropoda) were taken in most dredgings and appear to have been unaffected by the dosages of Toxaphene used in the experiments. Plankton In order to determine the effect of various Toxaphene concentrations on the more important planktonic groups, qualitative plankton sampling was carried out. The results are summarized in Table 5. Nine months after treatment, rotifers, flagellates and diatoms were present in all samples, with the exception of diatoms in Spectacle Lake (0.07 p.p.m.) (This exception could have been due to sampling error). It is concluded that rotifers, flagellates and diatoms are not adversely affected by the concentrations of Toxaphene used. Cladocerans were present in all of the lakes following treatment, although one of the two forms was absent in some instances. These results suggest that this group is unaffected by the concentrations of Toxaphene used in the experiment. It may be seen from Table 5 that occurrence of copepods was erratic. It is suggested that seasonal variation in abundance together with sampling error accounts for these results. In any case, copepods were taken at all concentrations, which indicates that this group is also tolerant of the Toxaphene concentrations used. The available evidence strongly suggests, therefore, that Toxaphene at the concentrations applied in this experiment, has no significant effect on the major planktonic forms.

Table V Plankton Samples Taken Before and After Toxaphene Treatment.

('Y' Indicates Presence of Particular Form, While "ô' Indicates its Absence).

Alleyne Taylor Gladstone Round Gallagher Spectacle Summit Lady King (0.01) Organisms (0.01) (0.03) (0.03) (0.07) (0.07) (0.10) (0.10) ABC ABC ABC ABC ABC ABC ABC ABC

Cladocera: Daphnia...... oxx xxx xox xxo xxo xxx zoo xxx Bosmina...... xxx xxx xxx xxx xxx xox xxx xxx Copepoda: Cyclops...... xxx xxo xxo xxx xxo xox xxo xxx Diaptomus..... xxo 000 000 xoo xxo xoo 000 xox Naupliuslarvae. xxx xxo xxx oox xxx xox xoo xxx Rotifera...... xxx xxx xxx xxx xxx xxx xxx xxx Mastigophora (Flagellates).... xxx xxx xxx xxx xxx xxx xxx xxx Bacillariaceae (Diatoms)...... xxx xxx xxx xxx xxx xxo xxx xxx

A-Samples taken before poisoning. B-Samples taken one month after poisoning. C-Samples taken nine months after poisoning. 46

Discussion

Toxaphene is an inexpensive and effective fish toxicant at all concentrations used in the experiment. Since complete kills were obtained even at the lowest concentration, additional treatments will be undertaken using concentrations of 0.005 p.p.m. and 0.0075 p.p.m. Although the lakes were toxic to fish nine months after treatment, it is anticipated that they will be clear within a few months (two of the lakes treated with 0.10 p.p.m. Toxaphene were successfully stocked with fish two years later). Undoubtedly lakes having a continuous outflow would clear much more rapidly.

Prolonged toxicity can be an advantage and disadvantage to the fisheries manager. It may put a lake out of production for a year or more, but it also assures a complete kill, a result which was not always achieved by the use of other fish toxins.

Results of these experiments do not support previous findings (Hemphill, 1954) in which it was stated that waters with a pH of 8.0 or higher treated with a concentration of 0.10 p.p.m. Toxaphene were planted with fish within four weeks. Perhaps the lake referred to had a high flushing rate or the concentration of 0.10 p.p.m. referred to the total amount of commercial product containing a low percentage of Toxaphene.

The British Columbia Research Council was approached concerning the factors which might control the rate of breakdown of Toxaphene in lakes and the possibility of applying chemical neutralizers. Apparently, Toxaphene breakdown is governed by a complexity of factors in which acidity, sunlight, temperature and bacterial action may be involved. It is doubtful that breakdown could be governed by a specific chemical compound.

In general, Toxaphene shows promise as an inexpensive and effective fish toxicant, its value increasing as more information is collected on minimum doses and length of toxicity. Acknowledgements The authors gratefully acknowledge the guidance of Dr. C. C. Lindsey in planning the experiment, and the ingenuity of Stuart B. Smith in the design and construction of the distribution equipment. Invaluable assistance was also provided by Depart• mental personnel especially John Balkwell in applying the chemical.

Literature Cited HEMPHILL, J. E. 1954. Toxaphene as a fish toxin. Prog. Fish. Cult. Vol. 16, No. 1, pp. 4142. 47

Abstract

Eight alkaline lakes in British Columbia were treated with Toxaphene at concentrations ranging from 0.10 p.p.m. to 0.01 p.p.m. Effects on fish, bottom fauna and plankton were evaluated. Most of the fish were killed in all concentrations within the first 120 hours. While a few fish were gill-netted, two to three weeks after treatment, none was netted in any of the lakes eight to nine months after poisoning. The lakes were still toxic to fish nine months after treatment as determined by live cage tests. Amphipods (shrimp) were eliminated at all concentrations and were still absent nine months after treatment. Dragonfly and damselfly nymphs were killed at a concentration of 0.03 p.p.m., while midge larvae were killed at concentrations of 0.03 p.p.m. and greater. No consistent adverse effects were noted with respect to plankton. Toxaphene was indicated as being an effective and economical fish toxicant; further work is being carried out with respect to more dilute concentrations and length of toxicity.

Review of Literature

"Freshwater Fishery Biology," by Karl F. Lagler, Wm. C. Brown Company, Publishers, Dubuque, Iowa, 1956; 434 pp., 214 figures. Cloth bound. $6.75. The new enlarged edition contains the same 25 chapters and six appendices as that of 1952 volume, but fortunately several errors and omissions in the previous edition were corrected. However some minor omissions still exist. For instance, on pages 390-402 in Appendix F, dealing with the economic classification of freshwater fishes, several are not considered as sport species. In reality the following are taken regularly on hook and line and provide pleasant recreation: smelt, brown bullhead, American eel, sauger and American shad. To this list should be added two species of Pacific salmon (Oncorhynchus), coho and chinook (or spring), which are taken on line not only for sport but for commercial purposes as well. In the same Appendix F, it is not very clear to us why some anadromous fishes such as sturgeon species (Acipenser) are included in the Est but others such as tomcod (Microgadus), or striped bass (Roccus saxatilis) are excluded. Moreover, several species of freshwater fishes are omitted from the list.

In chapters XVIII and XIX, dealing with Fishery surveys of iakes, some modem equipment such as echo sounding machines and bathy-thermographs should be mentioned. The section dealing with the Methods of marking of fish should be extended to include more recent types of tags, descriptions of which appeared after the publication of the review by Rounsefell and Kask (1946). In the section on Enumeration of fish eggs (pp. 106-111), the findings and techniques of more authors, especially recent ones, should be incorporated. A special section on stages of maturity of fish will be equally useful. Also, it would be advantageous for students to have the scientific term added after the common name under each figure of freshwater fishes, mentioned on pages 23-58.

The scientific nomenclature found in Appendix F (pp. 388-403) is up-to-date, including all new changes' in names of fishes, which were mostly suggested by Dr. R. M. Bailey.

Minor omissions, enumerated above, do not depreciate the real value of this book, already well-known from its earlier editions. Therefore, it is the reviewer's pleasure to recommend to everyone Dr. Lagler's volume, as a very useful and comprehensive treatise on freshwater fishery biology. Numerous illustrations, clear and comprehensive text, selected references at the end of each chapter, good paper all add to the attractiveness of the book. This new revised and enlarged editon is a "must" for every university library and biology teacher's desk-Vadim D. Vladykov, Department of Fisheries, 9.uebec.

' The reviewer, however, doea not agree with many changes propoeed by Dr. Bailey.

ISSUE TWENTY-FOUR FEBRUARY - - 1959

THE CANADIAN FISH CULTURIST

L1IMARY i•: ilE,RtES AND BiSLIOTHtQUE A PECHES ET

Published at Ottawa by The Department of Fisheries of Canada ISSUE TWENTY-FOUR FEBRUARY - - 1959

THE CANADIAN FISH CULTURIST

Published at Ottawa by The Department of Fisheries of Canada

THE @jJEEN'B PRINTER AND CONTROLLER OF STATIONERY OTTAWA, 1959

65542-3-1 CONTENTS

Aerial Chemical Control of Forest Insects with Reference to the Canadian Situation-F. E. WEBB ...... 3

Effects of Spruce Budworm Control on Salmon and Other Fishes in New Brunswick-MILES H. A. KEENLEYSIDE.... 17

Effects of Black-Headed Budworm Control on Salmon and Trout in British Columbia-R. A. CROUTER AND E. H. VERNON ...... 23

Toxicity of a DDT Forest Spray to Young Salmon-D. F. ALDERDICE and M. E. WORTHINGTON ...... 41

The first three of the above papers were presented at a symposium on "The Effects of Chemical Control of Forest Insects on Fish Stocks," held by the Committee on Biological Investigations of the Fisheries Research Board of Canada in Ottawa, January, 1958. A fourth paper submitted, "Effects of Spruce Budworm Control on Stream Insects in New Brunswick," by Dr. F. P. Ide, Department of Zoology, University of Toronto, was published in the Transactions of the American Fisheries Society, Vol. 86, 1956, pp. 208-219.

The Canadian Fish Culturist is published under the authority of the Minister by the Department of Fisheries of Canada as a means of providing a forum for free expression of opinion on Canadian fish culture. In the areas of fact and opinion alike, the responsibility for statements made in articles or letters rests entirely with the writers. Publication of any particular material does not necessarily imply that the Department share9 the views expressed. In issuing The Canadian Fish Culturist the Department of Fisheries is acting only as an instrument for assisting in the circulation of information and opinion among people in the fish culture field. Those who may wish to discuss articles which have been published in The Canadian Fish Culturist are encouraged to do so and space will be made available. Correspondence should be addressed to the DIRECTOR, INFORMATION AND EDUCA• TIONAL SERVICE, DEPARTMENT OF FISHERIES, OTTAWA, CANADA.

Published under Authority of HON. J. ANGUS MACLEAN, M.P., Minister of Fisheries Aerial Chemical Control of Forest Insects with Reference to the Canadian Situation by F. E. Webb

Science Service, Federal Departmcnt of Agriculture, Fredericton, TIB.

The Forest Insect Problem The value of the Canadian forestry resource is no more impressive than are the problems of protecting it. Of the various factors responsible for forest losses other than man, none exerts a more consistent drain on timber inventories than insect damage (5). A generalization about Canadian forests is that they are chiefly of the Boreal type. This means that huge areas are covered by relatively uniform stands, often of the same age-class, comprised of relatively few tree species and usually dominated by one or two of the conifers (7). This type is uniquely valuable as a source of pulp and paper, but is a relatively unstable ecological association and successional changes are commonly sudden and violent. On the other hand, the cataclysmic natural events that have played a part for centuries in the periodic replacement of successive crops- chiefly fire, blowdown, and insect outbreaks-generally tend to perpetuate the kind of forest that they ravage. Aspen, pine, and spruce are typical post-fire types; in eastern Canada conditions for the reproduction and growth of the ubiquitous balsam fir are never better than following the destruction of mature stands of the same species by the spruce budworm (6). These natural catastrophes played a dynamic role in succession before man began to exploit the forest but only since the rapid development of the Pulp and Paper Industry, with its full use of the forest growth, have they become serious economic problems. Some of the current problems with native insects arise partly from man's interference. Selective cutting of certain species or vigour classes may create a condition of higher susceptibility to insect and disease attack in the residual crop. From the point of view of the forest manager and the silviculturist, clear-cutting for pulpwood is often sound practice, yet it is the opinion of some that this is encouraging a higher proportion of balsam fir reproduction and intensifying the threat from the spruce budworm (16). The presence of large quantities of logging slash and logging injuries to residual trees sometimes increase the incidence of bark beetle and borer attack in surrounding uncut stands. Equally serious, numbers of foreign insects have been introduced over the years and some of these now pose major problems. ' In short, insect problems and the need for protection have tended to increase rapidly with increased accessibility and exploitation of our forests (13, 14).

65542-3-1 } The Place of Aerial Spraying in Forest Insect Control Most forest entomologists subscribe to the philosophy that the ideal in forest insect control is prevention rather than cure. The creation of resistant forests, however, takes time and can only keep pace with the development of forest management. It must be based on intensive ecological research translated into practical economic silviculture. Such research is being undertaken and forest management is becoming more intensive but the completely resistant forest is likely to remain an ideal, at least for many years to come (2). Biological control by the manipulation of parasites, predators, or disease organisms has been successful in some cases of introduced pests. The possibilities of their use against such native pests as the spruce budworm have been and are still being studied, but as yet without positive results. Meanwhile direct measures must be applied if severe outbreaks are to be prevented from seriously reducing the supply of raw material of the forest industries and at the same time disrupting management plans on which sustained yield and the ultimate production of more resistant forests depend. If, for instance, the spruce budworm kills large areas of forest, the succeeding stand will be even-aged and of high balsam content-difficult to manage for sustained yield, and highly susceptible to future outbreaks (12). The discovery of DDT and the development of aerial methods of application have made the use of insecticides against a number of forest insects practicable. This, combined with the increasing importance of the full use of our forest growth, explains the number of large-scale operations since the second World War (3). Forest entomologists are fully aware of certain limitations and possible dangers in I chemical methods of control. They have always been concerned with the population dynamics involved and have sought the aid of other workers in attempts to determine effects on fish and wildlife. However, when spraying presents the only means of protecting a forest from serious loss it becomes necessary to use it (1). At the same time this does not imply prolonged or continuous use of insecticides analogous to agricultural practices. In forestry it is not necessary to maintain insect populations at as low levels as in agriculture and large-scale use of aerial spraying in Canada has been limited to the prevention of tree mortality. These operations are being studied as closely as possible to determine their long-term effects on the population dynamics of the pest and associated insects (18). Aerial spraying operations may range in size from a few hundred acres as in the case of plots, or plantations, to several thousands of square miles as against the spruce budworm. In most cases they involve defoliating insects for which dosages of 1 lb. DDT/acre or less have become more or less standard. Important adverse effects against other forms of life have seldom been reported from small operations using dosages of this order, and recovery is generally rapid (4, 15). Serious effects are more likely from blanket coverage of extensive areas, particularly when treatment is repeated. For purposes of the present discussion, therefore, it will suffice to consider only the larger type of operations involving extensive forest areas. These are listed for Canada to date in Table 1. Since the spruce budworm has been far and away the most serious problem to date in Canada and because of my own experience with it, I propose in the following sections to refer in particular to the recent large-scale operations in eastern Canada. Table 1

History of Large-Scale Forest Spraying Operations in Canada. DDT-Oil Formulations Used in All Cases

Insect Year Province Dose: DDT Acres Sq. mi.

Western hemlock looper...... 1946 B.C. 1 lb./acre 12,000 19 False hemlock looper...... 1948 B.C. 1 lb./acre 11,000 17

Black-headed budworm ...... 1957 B.C. 1 lb./acre 156,000 240

Spruce budworm ...... 1945 Ont. 1 lb./acre 64,000 100 1946 Ont. 21b./acre 29,000 45 1952 N.B. 11b./acre 186,000 290 1953 N.B. '1 1b./acre 1,80C,000 2,800 1954 N.B. and Que. " 1,500,000 2,300 1955 N.B. and Que. " 2,20C,000 3,400 1956 N.B. and Que. " 2,400,000 3,750 1957 N.B. and Que. " 6,500,000 10,200

'Approximately 25 per cent of 1953 area treated twice.

Aerial Spraying with Particular Reference to the Spruce Budworm History of Outbreaks I Evidence can be found in forests of northeastern North America of outbreaks dating back over 150 years. Intervals between them have varied from about 70 years in northwestern Ontario to 35-40 years in New Brunswick. Severe infestations have been reported somewhere in eastern Canada practically every year for at least the last 40 years and an estimate has been made that between 1909 and 1946 outbreaks had killed some 250 million cords of balsam fir and spruce (5). Particularly widespread and severe outbreaks occurred between about 1909 and 1920 and from the late 1930's until the present time. In both these periods the outbreaks have shown a tendency to shift from west to east, beginning in northern Ontario and ending in the Atlantic regions of New Brunswick, Nova Scotia, and the Lower St. Lawrence-Gaspé regions of Quebec. At the present time outbreaks are at a relatively low ebb over much of Ontario and Quebec north of the St. Lawrence and are most severe in the Atlantic Maritime region. At the same time, however, trees are still dying from the attacks of the last 10 to 15 years over thousands of square miles in Ontario and Quebec and persistent pockets of infestation still occur in western Ontario that have been under more or less continuous attack for 15 years or longer.' In New Brunswick the current outbreak first became severe in 1949 and some idea of the rate of spread is shown in Table 2. It is also illustrated in Figures 1-3 showing the areas involved in three representative years, 1952, 1954, and 1957 respectively. It

'Annual Reports, Forest Insect Sun", Forest Biology Division, Science Service, Canada, Dept. of Agriculture, 1937 et seq. Table 2 Spraying Areas of Severe Attack by Spruce Budworm in New Brunswick and Areas Treated by Aerial

Areas-sq. mi.

Severe attack Sprayed

1949 ...... 200 0 1950 ...... 400 0 1951 ...... 2,200 0 1952 ...... 5,000 300 6% 1953 ...... 11,000 2,800 25 1954 .....:...... 13,000 1,800 14 1955 ...... 13,000 1,800 14 1956 ...... 16,000 3,100 19 1957 ...... 20,000 8,100 40 1958 ...... (3,900?)

Percentage of N.B. areas sprayed once: 13.7 twice: 55.5 three times: 28.3 four times: 2.5

Piouxs 1 9PfdICE BIAWORM NFESTATION AND AERUIL S4RAYNG N NEW EItl049WICK - 1964 - 01111KNAa 10 :wNE anoc sRArm rrA

should be pointed out that the table and maps refer only to areas of severe attack and that.there were also large areas of lighter attack in each of these years.

Brief History of the Aerial Method The spruce budworm was not the first forest insect to be treated from the air, but it is of interest that 1957 was the 30th anniversary of the first spruce budworm aerial control project ever attempted. That was in Nova Scotia in 1927. The conclusion reached in those early years was that while dusting was a useful method against insects such as hemlock looper, it was not likely to be highly successful against the budworm until better insecticides and equipment were developed. The advent of DDT oil sprays during the War answered many of the problems and a good deal of work was done in the immediate post-war years in both Canada and the United States to develop suitable application equipment and techniques. In 1949, the first of a series of large- scale annual operations against the spruce budworm was carried out in'the Douglas fir regions of Oregon and Washington and by 1955 these totalled nearly 4 million acres. Operations on a similar scale have been underway since 1953 in the Rocky Mountain regions from New Mexico to Idaho and Montana. The history of large-scale Canadian operations as summarized in Table 1 gives some idea ^of the relative importance of the spruce budworm problem as compared to other defoliators. The first budworm 8

FIGURE 3

operations, totalling some 93,000 acres, were carried out in 1945 and 1946 in the Lake Nipigon area of Ontario (10). The largest operations of all, however, have been underway since 1952 in New Brunswick and since 1954 in adjacent areas of Quebec. These now total about 14.5 million acres including respraying. This is an acreage in excess of the land area of several European countries including Switzerland. All the more remarkable is the fact that nearly half this was sprayed in a single year, 1957-nearly 6.5 million acres in the two provinces. New Brunswick's share alone amounted to 5.2 million acres.

Organization and Direction of New Brunswick-.6Zuebec Operations The organization and direction of the New Brunswick operations is the responsibility of a provincial Crown Company, Forest Protection Ltd. Mr. B. W. Flieger is the Manager. Shares in the Company are held by the four leading pulp and paper Companies in the province and by the New Brunswick Department of Lands and Mines. Costs are shared equally by Canada, New Brunswick, and industry. The Division of Forest Biology has been closely associated with the project since its beginning and is responsible for various aspects of technical advice and assistance. These include timing the spraying, assessing immediate results, and estimating hazard from surveys of infestation and defoliation. A programme of long-term investigation is 9 also under way to discover and evaluate ultimate effects on the epidemiology of the budworm and on the growth and survival of the forest (18).1 Quebec operations are carried out by Quebec Forest Industries Assn. representing landowners and leaseholders. Costs are shared equally between industry and the province. The operations in the two provinces are closely integrated. The same aerial fleet is employed for both jobs using identical timing, dosage, and techniques of application. A programme of assessment and investigation similar to that in New Brunswick is carried out by the Quebec Laboratory of this Division.

Spraying Poiicy The stage of an outbreak at which spraying is undertaken will vary according to the objectives. Some of the operations carried out in recent years against gypsy moth in the United States have treated very light infestations with the objective of limiting the spread of this new insect beyond an established barrier zone. These and some other examples are often referred to as "eradication programmes"-usually with dubious justification in the light of experience. Fortunately the gypsy moth is not important in Canada and there is yet no comparable problem from the foreign pests that do occur. The idea of spraying incipient outbreaks of native pests to prevent their rise and spread is an interesting possibility that deserves to be tested but it involves some difficult practical problems. The foci of outbreaks are not easily detected early in their development in forests the size of Canada's and are often very large in themselves. It will seldom, if ever, be possible to organize a large spraying operation against the same insect generation in which the outbreak is detected and difficult problems are involved in mapping the extent of spread of the next generation when this must be based on extended surveys of eggs or other minute immature forms. In addition, effective control at relatively low populations is likely to call for larger dosages than the minimum required against heavy infestations of defoliators owing to the greater concealment of larvae in the denser foliage of undamaged trees. The alternative to early spraying to control outbreak trends is to delay treatment until it is necessary to prevent tree damage. In dealing with the operations listed for Canada in Table 1 a consistent policy has been followed of recommending spraying only when a further year's defoliation would seriously threaten the life of trees. This is a conservative policy that avoids the uncertainties of spraying incipient outbreaks and gives natural control factors the best opportunity to exert a regulating effect before resorting to applied measures. However, it calls for a degree of judgement and a good knowledge of the ability of trees to withstand attack.

Techniques and Procedures FORMULATIONS AND DOSAGE: DDT is the most commonly-used insecticide again forest insects and has been used on all the large-scale aerial operations in Canada. This is because of its low cost and high toxicity, particularly against free- living defoliators. Its chief effect is as a contact poison but this is reinforced by

'Annual Summary Reports on these operations have been published in the following issues of the Bi, Monthly Progress Rept., Div. For. Biol., Dept. Agric. Canada. New Brunswick: 8(4): 10(1); 11(1); 12(2); 13(3). 19,uebcc: 11(1); 12(1); 13(3). 65542-3-2 stomach and residual poisoning. The most commonly-recommended dosage against forest defoliators in North America is 1 lb. DDT per acre, usually in one U.S. gallon of oil solvent. Only once in Canada has this been exceeded on a large scale-in northwestern Ontario in 1946 (10). It was used in the first New Brunswick operation of 1952. Since 1953, however, practically all the New Brunswick-Quebec spraying has been done at half this dosage by doubling the average distance between adjacent flights of the spray planes. With good timing and with some drift it has been found to give results nearly as good as the heavier dosage, especially when applied against conditions of very heavy infestation. In effect it has permitted the treatment of nearly double the area at the same cost and using the same aerial fleet and facilities. These advantages are considered adequate to make up for the somewhat lower average percentage of control that is obtained and for the slightly less uniform coverage that is inevitable. A limited amount of the New Brunswick spraying has involved two applications at the 2-lb. dosage two to three weeks apart. The earlier application provides good foliage protection while a high percentage of kill is ensured by the later spraying. This technique is considered more effective than the single application at the 1-lb. dosage and is the safest way of ensuring good results in cases of extreme hazard or where the values at stake are particularly high.

SPRAY EQUIPMENT AND CALIBRATION : The most widely-used dispersal apparatus is the boom and nozzle type, fed under pressure from a tank in the airplane fuselage. The pressure and the nozzles are adjusted to give the appropriate rate of flow and the proper range of droplet sizes. The optimum droplet size is the minimum that will reach the trees before evaporation. Too-fine sprays drift excessively and reduce the average deposit. Coarse sprays are wasteful, give poorer coverage, and may be more hazardous to aquatic life.

FLYING PROCEDURE AND SPRAY COVERAGE: A wide variety of aircraft types have been used from the smallest single-engined planes to multi-engined transports and converted bombers. A helicopter was used for the 1948 operation against false hemlock looper in British Columbia but their use has not otherwise been common owing to excessive cost. Light conventional aircraft are generally preferred owing to their relatively low cost of operation, greater manoeuverability at low levels and their availability. The most widely-used type is the Stearman bi-plane converted from the World War II elementary trainer. It has been used exclusively in New Brunswick and Quebec except for one de Havilland Beaver used in 1952. The aerial fleet has varied from 20 planes in 1952 to 200 in 1957. The procedure in most aerial spraying is to fly as close to the ground as is com- mensurate with safety and the problems of navigation. Forest spraying is commonly carried out at about 50 to 250 feet with the smaller airplane types depending mostly on topography. Pilots are assigned individual spray blocks of an average size of about 5,000 acres and coverage may be obtained by flying adjacent swaths from end to end over flat terrain or by "contouring" on slopes. In New Brunswick, the practice has been adopted of flying planes in pairs rather than singly. This has advantages in safety, simplifies the problems of supervision and control, and improves the chances of obtaining uniform coverage. Good coverage depends largely on the skill of the pilot in low-level navigation. With the most precise flying possible, however, uniform coverage would be unlikely without drift. For that reason low wind velocities are often considered advantageous and experienced pilots are skilled in allowing for and taking advantage of drift. Experi- ence has shown the necessity of isolating unsprayed check plots by at least one and preferably more miles from the nearest spray boundary and of spraying around such areas only with favourable winds. Although pilots are instructed to shut off the spray over open water, this seldom prevents some of the spray from entering streams and lakes. Water courses are some of the most easily-recognized topographical features and they very frequently form boundaries between spray blocks. For this reason coverage along some streams may average heavier than within the blocks. TIMING APPLICATIONS: A good deal of the success of aerial spraying depends on the timing of applications. The objectives can be twofold: to achieve the maximum possible reduction of the insect population and to prevent defoliation. The first objective is best achieved by spraying when most of the feeding is completed and larval exposure is at a maximum; the second calls for early spraying, usually with some sacrifice of percentage of kill. The policy in New Brunswick and Quebec has been to achieve the best compromise by bracketing the optimum period as effectively as possible (17). With the spruce hudworm this involves a period of about two weeks in a specific locality. Phenological differences within the areas under treatment vary as much as three weeks and advantage is taken of this to commence spraying in the more advanced southern areas and to shift operations farther north as conditions permit (19). More precise timing than that described here for the spruce budworm would be difficult to put into practice owing to the uncertainties of weather and the exigencies of large-scale operations. However, where more leeway may be possible, as in small operations or against other pests, consideration should be given to the possibilities of adjusting timing to reduce adverse effects against parasites and predators and other forms of life that may be particularly vulnerable during some part of the spraying period (2).

ASSESSING RESULTS AND EVALUATING HAZARD: In most operations spray coverage is checked on the ground by means of test cards or plates that are set out to sample the deposit in representative locations. This technique was used in New Brunswick and Quebec in the first.one or two years and it showed that good coverage of lethal deposits could be obtained over all types of terrain by good flying aided by drift. In the very large programmes that have followed, however, it has been necessary to place greater reliance on close aerial supervision of the spraying by inspection pilots. These men are experienced forest sprayers and they are also responsible for pre-flight briefing of spray pilots, and for judging the suitability of spraying weather. Ground assessment is still carried out on biological study areas and some use is being made of evidence of spray burn on hardwood and herbaceous foliage for extended spot checks.

Assessment of immediate biological results are necessary each year as a check on the efficacy of the toxicant and the effects of timing, weather, and techniques of application. Special problems are encountered in evaluating results and assessing hazard over very large areas such as those involved in New Brunswick and this has called for 65542-3-24 12

the development of suitable surveying and sampling methods. Aerial surveys have been used extensively every year to compare sprayed and unsprayed areas and to determine the boundaries of light, moderate, and severe attack throughout the Province. Large-scale ground surveys have been developed for two main purposes: (1) to measure immediate results of the spraying by comparing defoliation and post-spray budworm survival in sprayed and unsprayed areas, and (2) to measure egg populations of the next generation and combine this with the estimates of defoliation and damage for the purpose of gauging hazard. In both cases a system of sequential sampling has been developed that ensures the greatest economy of counting effort and takes the best advantage of limited time and staff. Studies of ultimate results are a long-term proposition that should preferably consider the whole ecosystem and the New Brunswick-Quebec operations provide a unique opportunity to undertake such studies. In view of the aggressiveness of this insect and the huge areas of susceptible forest this is probably the severest test of chemical control that has ever been made. Intensive studies of the epidemiology of this insect and the possibilities of reducing its threat by forest management were already well underway in New Brunswick when spraying began, and these studies provided the best possible basis upon which to extend the investigation to the effects of spraying (12). Comparisons are being made of population trends in sprayed and unsprayed areas and these are being related to measurements of regulating factors. The effects of spraying on the growth and survival of the trees are under study and this is contributing annually to our knowledge of the ability of different host species to sustain attack. Extended observations are made each year for the purpose of detecting any unusual occurrence of other pests and gross measurements are being made of the effects of spraying on other insects-both terrestrial and aquatic.

Results It is convenient to discuss results in terms of (1) insect populations including the pest and associated beneficial or destructive species; (2) the forest stand; and (3) fish and wildlife. A sound evaluation of all these effects is essential for an intelligent and balanced appraisal of the economic and biological values of forest spraying.

EFFECTS ON THE PEST AND ASSOCIATED INSECT POPULATIONS: Reports on the results of DDT applications against forest defoliators consistently show tremendous immediate reductions of most of the insects that are exposed at the time of spraying. Assessments of over 99 per cent control of the spruce budworm were obtained in 1952 in New Brunswick at the 1-lb. dose, and since 1953 at the 2-1b. dose reductions have averaged between 80 and 100 per cent as compared with unsprayed checks. Results that compare favorably to these have been reported from the United States'against this insect and from British Columbia against hemlock looper, and more recently, against black-headed budworm. Results against the pest may also be judged by comparing sprayed and unsprayed populations in the next generation. In New Brunswick and Quebec rapid reinfestation of sprayed areas is a particularly troublesome aspect of the problem. It results not only from a resurgence of residual populations but also from large-scale reinvasions of 13 moths from surrounding unsprayed areas. It is for this reason that most of the areas involved so far have been sprayed twice and some areas three and four times (Table 2, Figure 4). Parasites and predators of the pest that are vulnerable at the time of spraying are often much reduced. Some of the more important parasites of the spruce budworm do not seem to be affected as severely as the host, however, and proportionately heavier parasitism in sprayed areas has been reported from both the United States and Canada

. ^•...,•.••`_^._ -.',

SPRUCE BUDWORM SPRAYING PROGRAMME - 1952-1957

SPRAYED ONCE

' TWICE

' TFHEE TIMES

FOUR TIMES

ATLANTIC SAl#MON STREAMS

FIGURE 4

(11). New Brunswick studies also show insect predators to be capable of proportionately better survival. Much more remains to be learned about the effects of the treatment on the density-dependant relationships of surviving populations. Both terrestrial and aquatic insects, other than the pest and its parasites and predators, are temporarily much reduced and some species are affected more severely than others. Excellent control of black flies and mosquitoes usually results. Most terrestrial groups, however, appear to be capable of quick recovery and total numbers of flying insects are often at least as great as in unsprayed areas later in the same season. An interesting study in the United States showed that a sprayed hardwood stand 14

supported a larger general insect population than an unsprayed stand subjected to exposure by defoliation (8). More work is needed to distinguish between species that are severely affected and those that are not, and to assess the changes that undoubtedly occur in the complex of post-spray populations. There are no signs yet in Canadian operations that other destructive species are being encouraged. A recent report of widespread mite damage in forests sprayed against spruce budworm in Montana and Idaho, however, emphasizes the fact that this remains a possible hazard. The phenomenon of resistance to DDT reported for a number of agricultural pests has not yet been encountered in forest defoliators. While this also remains a possibility, its development will probably require the treatment of many more successive generations of the pest than has occurred so far.

EFFECTS ON THE FOREST: From the point of view of the forester, the best evaluation of the results of spraying is in terms of its effect on the forest stand. How successfully was the objective achieved of preventing mortality? What has been the effect on the growth capacity of the trees, and what will be the ultimate silvicultural and economical benefits? As in the case of effects on the pest, a full appraisal calls for assessment of both immediate and long-term effects. Most of the reports available to date from Canada and the United States tell of success in preventing mortality but show little or no concern with the preservation of annual growth. Some change will undoubtedly occur in this emphasis with the fuller utilization and more intensive management of the forest crop. With each succeeding year of operations it is becoming increasingly difficult to generalize about the condition of sprayed forests in New Brunswick. Conditions ranging from excellent growth recovery to outright mortality are present and often are found within a short distance of one another. There can be little doubt, however, that balsam fir mortality has been substantially forestalled or delayed and that recovery potential remains good over most of the five million acres that are now involved. This is evident from unsprayed check areas where after eight years of severe attack the merchantable component of stands is dead or dying and stocking in younger stands is being depleted. Mortality in sprayed areas is still largely confined to small, scattered patches and chances are good that many of these will be salvaged. As the outbreak continues in New Brunswick, the problem of assessing the need for respraying is more and more becoming one of estimating the ability of trees to withstand successive attacks and due allowance must be made for their lowered vitality. It now seems evident that spraying may be necessary every other year or, at the most, every third year as long as the outbreak persists and if trees are to be kept alive. EFFECTS ON FISH AND WILDLIFE: Studies of effects against fish are not a function of the Division of Forest Biology and I shall not presume to review this field here. I might say, however, we follow with keen interest the work of Fisheries Research Board in connection with Atlantic salmon studies and we have welcomed the opportunity to supplement the work of Dr. Ide on the Miramichi (9) in a small way by surveying aquatic insect populations in some sprayed and unsprayed streams in another part of the sprayed area. There has been relatively little intensive work done in New Brunswick or elsewhere in Canada on effects against wildlife. The concensus, both here and in the United States, however, suggests that dosages customarily recommended against forest defoliators are not likely to have serious adverse effects directly or indirectly against warm-blooded animals. This opinion is supported in an excellent review of the problem by Rudd and Genelly published recently as a Bulletin of the Game Management Branch of Department of Fish and Game, California (15). Some of their recommendations dealing with the use of pesticides in agriculture as well as in forestry may be of interest: (1) The most selective (pesticide) should be used. (2) Application should be in strict accordance with prescribed recommendations, should be used only where needed, and should be restricted to the minimum effective amount. (3) Greater concern should be accorded to widespread application than to local use. Serious effects are more likely to result from treatments over large areas. (4) Field biologists should participate in tlie formulation of use procedures and should be consulted at any time when pesticides are applied over a wide area. (5) Cultural or biological means of control should be substituted for chemical wherever possible. (6) For the greatest safety, "insurance" or routine applications of pesticides should be avoided. Concluding Remarks Nowhere else in the world has the development of large-scale aerial chemical control against forest insects been a more logical development than in Canada. In the few years since the method became economically and operationally feasible, it has taken its place as an adjunct to the management of large, relatively inaccessible forests where other methods are yet ineffective. Further improvements in technique undoubtedly are possible and research on insecticides may be expected to uncover new compounds and formulations that will be more effective, or more selective in their action than DDT. It seems safe to say that the method will continue to be used with improved efficiency against an increasing number of pests. . No large-scale forest spraying project should be undertaken without a full realization of its possible adverse consequences. However, it is not sufficient to oppose the method on the basis of some vague notion of "disturbing the balance of nature." A more precise knowledge must be gained of the effect of introducing the new and powerful control factor into an already complex biological system. The full poten- tialities of the method can only be determined from tests on a sufficient sc4le to achieve their purpose, and operations such as those of recent years in Canada provide the necessary opportunities. The biologist's task is to foster the idea that the problems of pest control are primarily ecological and, by means of properly integrated studies of the whole ecosystem, to ensure the safest and most intelligent use of direct control measures. Literature Cited (1) BALCH, R. E. (1953) For. Chron., 29:6-13. (2) BALCH, R. E. (1958) Ann. Rev. Ent. 3:449-468. (3) BALCH, R. E., WEBB, F. E. and FETTES, J. J. (1955-56) Forestry Abstracts, Leading Article 23 16(4); 17(1 and 2). (4) BROWN, A. W. A. (1951) Insect Control by Chemicals. John Wiley & Sons, New York. (5) DE GRYSE, J. J. (1947) Canada Year Book 1947 (reprint, 14 pp.). (6) FLEIGER, B. W. (1953) Rept. Ent. Soc. Ont. 84:9-16. (7) HALLIDAY, W. E. D. (1937) Can. Dep. Mines and Resources. For. Serv. Bull. No. 89. (8) HOFFMAN, C. H. and LINDUSKA, J. P. (1949) Scientific Monthly, 69:104-14. (9) IDE, F. P. (1957) Trans. Amer. Fisheries Soc., 1956, 86:208-19. (10) JOHNSTON, R. N., LANGFORD, R. R., SAVAGE, S., LAGIER, E. B. S., HOPE, C. E., MacKIE, R. E., BROWN, N. R., ADDISON, P., STEWART, K. E., and SPEIRS, J. M. (1949) Biol. Bull. 2, Div. Res., Ont. Dep. Lands and Forests. (11) MACDONALD D. R. (In Press) Cart. Entomologist. (12) MORRIS; R. F., MILIER, C. A., GREENBANK, D. O., and MOTT, D. G. (1958) Proc. 10th Int. Congr. Ent., 1956, 4:137-148. (13) PREBBLE, M. L. (1954) Repy. 6th Commonwealth Ent. Conf., 1954, pp. 206224. (14) PREBBLE, M. L., BALCH, R. E., BARTER, G. W., MORRIS, R. F., LEJEUNE, R. R., HOPPING, G. R., RICHMOND, H. A., and KINGHORN, J. M., For. Chron. 27:6-37. 1951. (15) RUDD, R. L. and GENELLY, R. E. (1956) Game Bull. No. 7. State of Calif., Dept. Fish and Game, Game Manag. Branch. (16) TOTHILL, J. D. (1958) Proc. 10th int. Congr. Ent., 1956, 4:525-531. (17) WEBB, F. E. (1955) For. Chron. 31:342-52. (18) WEBB, F. E. (1958) Roc. 10th Int. Congr. Ent., 1956, 4:303-316. (19) WEBB, F. E. (1958) Pulp Paper Mag. Can. 59(C). Effects of Spruce Budworm Control on Salmon and Other Fishes in New Brunswick by Miles H. A. Keenleyside Fisheries Research Board of Canada, Biological Station, St. Andrews, N.B.

Each spring since 1952 large forest areas of northern New Brunswick have been sprayed with DDT in efforts to control a spruce budworm epidemic. The DDT has been mixed with a special solvent oil and sprayed from airplanes at a concentration of one half pound per acre (one pound per acre in 1952). Some areas have been resprayed at intervals of one, two or three years. In spite of these extensive and costly control measures, the budworm population continued to expand until in 1957 over five million acres of New Brunswick woodland was sprayed, much of it for the second or third time. In the face of such widespread and repeated applications of DDT many biologists, sportsmen and naturalists have become increasingly worried about the side effects of the insecticide on fish and wildlife in the budworm-infested areas. Fisheries workers have been especially concerned because many species of fish are known to be extremely sensitive to DDT. Fortunately, a unique opportunity has arisen to study the effects of this DDT spraying on the fish populations of some New Brunswick rivers. Since 1950 the Fisheries Research Board of Canada, through its Biological Station at St. Andrews, has carried out an annual fall census of the young salmon in several branches of the Mirami- chi River system of central New Brunswick. This programme has involved sampling the fish populations by seining with electrofishing at the same localities each year. Censusing began on the Northwest Miramichi River in 1950, was extended to the Dungarvon River in 1952 and to the Renous and Cains Rivers in 1955. In 1957 the work on the Dungarvon River was discontinued to enable censusing to begin on the Tobique River, the most important salmon angling tributary of the St. John River. A total of 34 census stations is involved : ten on the Northwest Miramichi, six on the Dungarvon, five on the Cains, three on the Renous and ten on the Tobique.

Effects of DDT spraying on young salmon The relative abundance of young salmon in the Northwest Miramichi branch of the Miramichi River each year from 1950 to 1957 is shown in Figure 1. Data from the first year of censusing (1957) on the Tobique River are included in the figure for comparison. All data are expressed as numbers of fish per 100 square yards of river bottom. The fish are separated into three groups on the basis of size. The smallest fish are fry or underyearlings. Parr less than 10 cm. total length are classid as "small parr"; those over 10 cm. as "large parr". In general, these three size groups correspond to the first three years of life. In the Miramichi River system most young salmon go to sea as smolts at the beginning of their fourth summer.

65542-3-3 18

Populations of all three size groups of young salmon in the Northwest Miramichi increased gradually from 1950 to 1953 (Figure 1). This increase is probably related to the experimental merganser control which has been carried out on the river since 1950. Control of American mergansers on the Pollett River, N.B., between 1947 and 1950 resulted in a marked increase in the number of smolts leaving the river (Elson, 1950). Following this evidence that reduction of merganser populations benefited young salmon, merganser control was extended to the Northwest Miramichi on an experimental basis in 1950. The experiment was designed to determine whether control of mergansers would eventually increase the numbers of adult salmon from this river taken in the sport and commercial fisheries.

1954 '58 '57 '57 V V V V FRY

p7TJTa ED 1:1 SMALL PARR

Z 15

^ IS LARGE PARR < ® 1950 '51 '52 '53 '54 '55 '56 'Si I '57 NORTHWEST MIRAMICHI TOBIQUE

Ficuaa 1. Abundance of young salmon in the Northwest Miramichi and Tobique Rivers. Solid triangles indicate years when DDT was sprayed in the area. Cross-hatching indicates fish found in sprayed parts of rivers.

In 1954 all stations sampled on the Northwest Miramichi were within the area sprayed with DDT for the first time. The drastic effects of the.spray are clearly seen in Figure 1. Not one fry was found that year. Small and large parr were also reduced but to a lesser extent than the fry. Direct evidence of the harmful effects of the DDT was obtained in 1954 by holding salmon parr in cages in several parts of the Northwest Miramichi. From 63 per cent to 91 per cent of those held within the spray area were dead in three weeks, while only 2 per cent died in an unsprayed control stream during the same period (Kerswill and Elson, 1955). The extreme sensitivity of young salmon to DDT was also demonstrated by laboratory bio-assays carried out in the summer of 1957. The median tolerance limit (concentration killing one half a sample of fish) for small parr in DDT was 0.049 ppm. for a 24-hour period, and 0.047 ppm. for 48 hours (Keenleyside, -1958). No spraying was done in the Northwest Miramichi region in 1955 and fry were unusually abundant that year (Figure 1). This heavy fry population may have been due to several factors: (1) less competition for food and space with salmon parr, many of which were killed by the 1954 spraying; (2) great abundance of aquatic Diptera larvae which, in contrast to many of the larger aquatic insect larvae, survive the budworm spraying and which form an important part of the diet of salmon fry in this area; (3) unusually successful spawning of adult salmon in the Miramichi area in 1954. Small and large parr were scarce in 1955, due to the lethal effect of the 1954 spraying. Those small parr found in 1955 may have been fry which were missed in the 1954 censusing, or they may have been small parr in 1954 which failed to reach large parr size by the fall of 1955 when censusing was done. The aquatic stages of many caddisflies, stoneflies and mayflies are greatly reduced by the DDT spraying (Ide, 1957). Since these insects form the major part of the diet of small and large salmon parr in the Miramichi area, a reduction in their numbers may be reflected in slower growth of the fish. In 1956 and 1957 only the lower part of the Northwest Miramichi River was within the spray area. In both years approximately the same extent of the headwaters was free of spray. The bars on Figure 1 for 1956 and 1957 are partly cross-hatched and partly open, to represent the proportions of fish found in sprayed and unsprayed parts of the river respectively. Clearly the best fry survival occurred in the unsprayed headwaters. In 1956 small parr were abundant, even in the sprayed section of the river. These were survivors of the very large 1955 fry class. Many of them were in poor condition in the fall of 1956, due to the scarcity of large aquatic insects on which they usually feed. Small parr were surprisingly abundant in the river in 1957. Most of them were found in the unsprayed headwaters, where they were more numerous than the fry of 1956. Some of these fish were probably small parr the year before which had not reached large parr size by 1957. Large parr were scarce in the North- west Miramichi in 1956 as a result of both the 1954 and 1956 sprayings. In 1957 they were numerous due to growth of the big 1955 fry class, some of which should have reached large parr size by 1957. All 10 census stations on the Tobique River were within the area sprayed in 1957. Some of them were also sprayed in 1953, 1955 or 1956. Figure 1 shows that young Tobique salmon were found in moderate numbers only in 1957. The figure for fry is misleading, since all fry found in. the river were taken at one station, where hatchery underyearlings were planted shortly before censusing. No fry were taken at the other nine stations. All the data obtained from population assessments on four branches of the Miramichi River since 1950 are summarized in Table I. The table includes the Northwest Miramichi data of Figure 1, but not the Tobique data. Fluctuations in abundance of young salmon over this larger area have been similar to those on the Northwest Miramichi branch. In years when DDT was sprayed populations of young salmon in streams within sprayed areas have been reduced. This reduction has been greatest among the fry. One year later, with no further spraying in the same areas, fry have been very abundant, while small and large parr have been scarce. In areas where no respraying was done for two or three years fry populations were similar to pre-spray levels. Parr, however, were more numerous than before, reflecting growth 65542-3--3} Table I

Salmon abundance in the Miramichi River since 1950

Number of Average no. fish per 100 sq. yds. stations studied Small parr Large parr

Before Spraying ......

Same Stations After Spraying Year of spraying ...... 2.5 9.4 3.7

( 1 year later . . . 50.8 0.6 2.8

No Further Spraying .... { 2yearslater.. 24.2 43.9 2.6

I^ 3 yearslater.. 27.8 48.4 29.7 and survival of fry of the first post-spray year. Also, slower growth of small and large parr, due to reduction of their food by spraying, probably results in some parr being included in the same size group for two consecutive years. It should be emphasized that abundant post-spray populations of small and large parr have been found at only three census stations, where spraying has not been repeated for an interval of three years (bottom line of Table I). These three stations are in the headwaters of the Northwest Miramichi River, which is excellent parr-rearing water. Most of the Miramichi watershed within the range of the present spruce budworm outbreak has been sprayed on successive or alternate years since 1954. Since most young salmon in this area spend three years in fresh water before going to sea, respraying with DDT at intervals of less than three years can affect the same year- classes more than once and thus greatly reduce the potential salmon production of the affected rivers.

Effects of DDT spraying on other fishes During the annual censusing of young salmon several other species of fish are regularly caught. . Changes in abundance of these species have also occurred following DDT spraying. The relative abundance of brook trout (Salvelinus fontinalis), eels (Anguilla rostrata) and four species of "minnows" in the Northwest Miramichi River since 1953 is shown in Figure 2. Under "minnows" are included four species of Cyprinidae, the black-nosed dace (Rhinichthys atratulus), the common shiner (Notropis cornutus), the fallfish (Semotilus carporalis) and the chub (Couesius plumbeus). Trout are found mainly in the headwaters of the Northwest Miramichi, a section of the river that was sprayed in 1954 only. They were scarce in 1954, recovered gradually until by 1956 they were more numerous than before spraying, and by 1957 were present in about average numbers again (Figure 2). Since the fishing gear used in censusing does not capture trout underyearlings efficiently, only data for yearling and older trout are included in the figure. Fluctuations among these age-groups are similar to those among salmon parr in this headwater section of the river. They are most abundant two years after spraying, when spraying is not repeated. Eels are present in moderate numbers throughout the Northwest Miramichi. They appear to have been severely affected by the DDT spraying (Figure 2). Their numbers were reduced in 1954 and in 1956, and by 1957 none at all were found. Minnows are found only in the lower part of the Northwest Miramichi, where spraying occurred in 1954, 1956 and 1957. They have remained abundant throughout this period and were particularly numerous in 1956 (Figure 2). The relative proportions of the four species in this river have been approximately the same during this period. Dace are about 10 times as abundant as each of the other species.

TROUT

. MINNOWSr

1953 'SI 'SS

FIGURE 2. Abundance of several species of fish in the Northuust Miramichi Rivtr. Symbols as on figure 1.

Summary Several species of fish in New Brunswick rivers have been adversely affected by the widespread aerial application of DDT to the forests of the province as a control agent for spruce budworm. Young salmon, brook trout and eels are reduced in number following spraying. Four species of cyprinids are relatively uriaffected by the spray. Salmon populations appear to recover if spraying is not repeated over the same are-as for an interval of at least three years. However, much of the budworm- infested woodland has been resprayed at intervals of less than three years. Under these conditions one age-class of salmon can be affeçted more than once and smolt production from rivers in the area will be seriously curtailed. 22

Acknowledgements The raw field data collected annually by the census crews has been processed and tabulated by Dr. P. F. Elson. His assistance in the preparation of this paper is gratefully acknowledged. Literature Cited

ELSON, P. F. 1950. Increasing salmon stocks by control of mergansers and ltingfishers. Fish. Res. Bd. Canada, Atlantic Prog. Rept. No. 51, pp. 12-16. KEENLEYSIDE, M. H. A. 1958. Comparative effects of the insecticides DDT and Malathion on young Atlantic salmon. Fish. Res. Bd. Canada, Atl. Prog. Rept. No. 69, pp. 3.6. t KERSWILL, C. J. and P. F. ELSON. 1955. Preliminary observations on effects of 1954 DDT spraying on Miramichi salmon stocks. Fish. Res. Bd. Canada, Atlantic Prog. Rept. No. 62, pp. 17-24. IDE, F. P. 1957. Efject of forest spraying with DDT on aquatic insects of salmon streams. Trans. Amer. Fish. Soc., 86: 208-219. Effects of Black-Headed Budworm Control on Salmon and Trout in British Columbia by

R. A. Crouter, t Dcpanment of Fisheries of Canada and E. H. Vernon, British Columbia Game Commission

Introduction An infestation of black-headed budworm, Acleris variana (Fern.), on Northern Vancouver Island had been under study by the Forest Biology Division of the Depart- ment of Agriculture Science Service since 1954. Surveys conducted in the fall of 1955 indicated that the infestation had reached high hazard proportions in the Port McNeill-Port Hardy-Quatsino Sound area. Defoliation within this area in 1956 was severe for the second consecutive year although a survey conducted in the fall of 1956 indicated that the egg count had declined 63 per cent from 1955. Scientists of the Forest Biology Division stated however, that unless over-wintering conditions in 1956 caused a collapse of the infestation, another defoliation would occur in 1957. They stated also that another defoliation, even of medium proportions, might cause tree mortality and top-killing within the high hazard area. For these reasons, it was recommended that consideration be given to carrying out control operations in the spring of 1957. In June of 1956, an aerial control experiment was conducted within the affected area to determine the dosage and timing of effective DDT treatment. To consider all aspects of this infestation, the British Columbia Loggers Associa- tion formed a Pest Control Committee in 1956. This committee met periodically with the Forest Biology Division and were kept well informed on the status of the budworm infestation. On the basis of the Forest Biology Division recommendation that consideration be given to carrying out control operations in 1957, the committee investigated the factors and requirements associated with the initiation of such a programme. Assurance of financial assistance was received from both the federal and provincial governments. The committee then instituted a plan to conduct an aerial spray programme in the high hazard area during the early summer of 1957. The final decision to spray was to be dependent upon the findings of a survey conducted by the Forest Biology Division in the spring of 1957, which would show if the irifestation had significantly lessened over the winter. This survey indicated that severe de5oliation would again occur in 1957 and plans for the aerial control programme were finalized.

23 6 The possibility that an aerial spray program might be conducted was first brought to the attention of the Department of Fisheries as a result of a report on the preliminary spraying experiment in June of 1956. Following discussion of the matter with the Forest Biology Division, technical personnel of the Department and the B.C. Game Commission were invited to attend meetings of the Pest Control Committee to discuss fisheries problems that might be associated with the proposed spray programme. Both groups remained in close contact with the committee throughout all stages of the operation. The effects of DDT spraying on fish and fish-food in other regions were closely studied and these were reviewed with the Pest Control Committee. Information on the location of salmon and trout streams within the proposed spray area was presented along with estimates on the size of salmon populations that might be affected. Because of the extreme hazard of aerial DDT spraying to fish populations, it was pointed out to the committee that unless certain major fish producing areas were eliminated from the spray area, there was every probability that extensive fish mortalities would occur. The committee, although sympathetic to the problem, advised that if these areas were eliminated, the control programme would be rendered relatively ineffective and as a consequence they could not agree with the suggested modification. Results from the experimental spraying indicated that a dosage of one-half pound per acre might be effective and the fisheries groups proposed that the recommended dosage of one pound per acre be reduced. The Forest Biology Division indicated however, that there was insufficient evidence that the lower dosage was effective and accordingly could not take the responsibility of recommending a reduction in the dosage. In order to reduce to a minimum the probability of damaging fish populations, the committee did agree however to the following proposals: (1) Streams would not be used as boundaries for spray plots. This would prevent streams from receiving a double dosage of spray as a result of overlapping. (2) Pilots would spray parallel to the course of the major streams, keeping one swath width away. (3) Spray would be shut off where it was necessary to cross streams. When the final decision to spray was made, a co-operative program was initiated by the Department of Fisheries and the B.C. Game Commission to assess any damage that aerial spraying might cause to the fish or fish-food populations. The Committee agreed that such an investigation would be particularly desirable not only in connection with the proposed programme but also for future reference.

METHOD Description of Spray Programme During the period June 10 to June 20 inclusive, 155,000 acres of forest were treated as outlined in Figure 1. Spraying was to commence when the budworm population had developed to the point where the majority of larvae were in the second instar, and with the first and second instars about equally represented. Insect development varied with altitude however and consequently the lower elevations were sprayed first and higher elevations later. Figure 1-Effects of Black-Headed Budworm Control on Salmon and Trout in British Columbia Four TBM Grumman Avenger aircraft, operating under contract with Skyway Air Services, conducted the spraying. These aircraft had a carrying capacity of 780 gallons of insecticide and operated at a speed of 150 miles per hour. The formulation constituted one pound of DDT dissolved in a wood penetrating oil with an emulsifier added and blended to one U.S. gallon with diesel oil. The rate of spraying was one U.S. gallon per acre. In addition to the three preventative measures agreed upon to reduce the probability of severe fish mortality, the Pest Control Committee instituted the following: (1) The concentration of DDT in the spray formulation was reduced to one-half pound per acre in the area between Englewood and the Nimpkish river, which contains a high proportion of second growth timber. (2) Spraying on the lower watershed of the Keogh river, which consists partially of cedar swamp, was confined to stands of hemlock. (3) The south-east side of Colonial Creek, which lies very near the border of the spray area, was not treated. In order to supervise the control programme, the spray company operated an observation aircraft. The Committee agreed to fly a fisheries observer in this aircraft and an arrangement was drawn up whereby either a Fisheries Officer or an Officer of the B.C. Game Commission was available for every flight. These observations were conducted to determine if it were possible to carry out the suggested preventative measures and also because it was felt that having the officers observe would provide an added incentive for the pilots to effect those measures.

The Assessment Programme Within the spray area, all five species of Pacific salmon, plus rainbow, cutthroat, and steelhead trout are indigenous. By spray date however, the seaward migration of pink, chum, and spring salmon fry and also that of coho, sockeye and steelhead smolts should have been completed as well as the movement of sockeye fry from stream to lake. The timing of the programme was such that adult salmon would not be present in the streams. Coho fry, trout, and juvenile stages of steelhead were the populations mainly affected by spraying. The treatment area was situated in an isolated region and access to many of the streams was difficult. There was, however, a system of private and public roads between Port McNeill, Port Hardy and Coal Harbour and in that area, six stations were set up for intensive studies. These were: station I on the Cluxewe river, stations II, III, and VI on the Keogh river, station V on Unnamed creek, a tributary of the Quatse river and station IV, a control on the Quatse river. (See Figure 1). At each of these stations three live pens were established and in these, coho fry, hatchery fingerling trout, and native steelhead smolts were retained. Mortality, water temperatures, and water levels were recorded daily and water samples were taken at stations II, III, and VI. Bottom samples were taken, using a Surber stream bottom sampler. Samples were taken prior to and during the spray programme and again in late July and in October. In addition, seining was conducted at station VI on the Keogh river. Spray cards supplied by the Forest Biology Division were set out at each station. Nine stations, consisting of single live pens containing coho fry were established throughout the more remote regions. The mortality after spraying was recorded for comparison with that measured in the intensive study area. In October, bottom samples were taken at each of these stations and at five additional "outside" streams. The water samples taken at the three stations were sent to the Nanaimo Biological station of the Fisheries Research Board of Canada, and were analyzed for DDT content. In addition, a series of bio-assay tests were made at the Biological Station, to determine the toxicity of both the spray formulation and its component parts to coho fry.

Results The total loss of fish in the pens was recorded and corrected for natural mortality to give an estimate of the initial kill directly attributable to DDT. Three control

CUMULATIVE PERCENTAGE OF NATURAL MORTALITY FOR COHO FRY ----- FOR HATCHERY FINGERLINGS

r i i ^. _.-^ ---- :: y ......

^ ... .

10 15 20 Days Confined to CsQes

FIGURE Z. The average natural mortality for fry and fingerlings as at stations I, II, III, V, an.i VI. The last measure of fingerling mortality represents only the loss at Station I. 27

pens were set up outside the spray area; one on the Quatse river, another on the upper portion of the Keogh river and a third on Colony creek. Water temperatures rose to lethal levels at station IV on the Quatse river causing a high mortality, and pen number III on the upper Keogh river was inadvertently sprayed. The mortality in the remaining control pen, station X, was low but it was felt that one pen could not provide sufficient control data. The natural mortality was calculated by grouping the fish stocked in pens I, II, III, V and VI, which were inspected daily, into one sample and calculating a total cumulative percentage mortality up until spray date. The initial total and the mortality for the individual streams were removed from the group totals as each was sprayed. The calculated percentage cumulative natural mortality for coho fry and hatchery fingerlings is shown graphically in figure 2. The initial DDT mortality figures for stations that were inspected daily have been corrected by taking the total loss after spraying and subtracting the expected natural mortality for that period as inferred from the graph. The figure for those streams that were checked only before and after, were corrected first by subtracting the inferred mortality for the period between the last observation after stocking and the spray date to give an estimate of the number of penned fish affected by spraying. The difference between that figure and the final number remaining alive was corrected again for expected natural loss for the period after spraying. The corrected DDT mortalities, listed in table I, as measured by the pens range from 0 to over 90 percent. The variation reflects a number of factors such as visibility of streams from the air, drift of spray, forest cover, proportion of the stream sprayed, concentration of DDT used and meteorological conditions.

(1) The Cluxewe River-Station I There was no measurable mortality of either coho fry or hatchery fingerlings attributable to spraying at station I. The portion of this watershed lying within the spray area was extensively logged and spraying was conducted on scattered patches of timber only. The spray formulation in this area did not contain an emulsifier but the scattered spraying probably accounts for the lack of mortality.

(2) The Keogh River-Stations II, III, VI The Keogh river with an estimated escapement of 40,000 coho in 1956, was of vital concern. Three live pens were established on this river, one near the mouth (station VI), another eight miles upstream (station II), and a third (station III), six miles above that. Table I lists the initial mortalities of coho fry as 3.0 per cent at the lower station, 87.2 at the middle and 91.2 at the upper. The hatchery fingerlings suffered losses of 9.5, 84.7 and 51.7 per cent at these respective stations. Subsequent surveys in July and October however, indicated that salmonids were almost absent from the system. The Keogh watershed was sprayed on June 10 and 12. Seine hauls at the lower station on June 17 and June 21 netted an average of 91 and 75 coho fry respectively, four hauls in late July averaged 8 fry and tour hauls at the same location in October netted no fish. At the middle station, a series of 25 hauls prior to spraying Table I

The estimated 1956 escapement to streams within the spray area and the mortalities of coho fry and hatchery fingerlings in pens at the assessment stations.

Corrected Mortality 1956 Coho Station Escapement Coho Fry ^o

Cluxewe R ...... I 2-5000 0 Keogh R ...... 40,000 II 87.2 84.7 III 91.3 51.7 VI 3.0 9.5 Quatse R ...... IV Control Unnamed Cr ...... V 500-1000 0 0 Coho Cr ...... Vil 2-5000 29.7 Nimpkish R ...... VIII 95.9 Waukaas R ...... 1% 500-1000 93.2 Colony Cr...... X Control Benson R ...... XI 1-2000 72.0 Colonial R ...... XII 500 94.9 Ingersoll R ...... XIII 500 0 Klaskish R ...... XI V 50C-1000 83.8 Mills Cr ...... 500-1000 Hyde Cr ...... 50-100 Lagoon Cr ...... 50-100 Rupert R ...... 3-500 Kwokwesta ...... 50 Cayeghle ...... 50 East Cr ...... 50-100 Teeta Cr ...... Cayuse Cr ......

averaged 34 coho fry, six hauls immediately after averaged 3.8 and three hauls in October netted nothing. Seining was not conducted at the upper station but only three fry were observed in 100 yards of stream on October 9. Coho fry appeared quite numerous at this time in each of the outside streams sampled for bottom organisms. Seining was conducted on two of thcse streams and on one, the Quatse river, six seine hauls averaged 10 coho fry and one trout fry per set. On the other, the Koprino river, four hauls averaged 17 coho fry. The initial loss in the Keogh, although quite heavy at the upper and middle stations was negligible at the lower site. The lower four miles of the watershed were only partially sprayed but a heavy oil slick was noted at the pen site. It was felt at the time that the DDT must either have settled out or have been absorbed in some way and that the fry in this area had been unaffected. Analysis of water samples later showed however that toxic concentrations of DDT were present at this site for more than 48 hours after spraying and subsequent surveys indicated a near complete absence of coho fry throughout the system by October. There is a possibility that the coho fry were displaced downstream and eventually to the sea in search of food. Since there is no evidence that fry which enter salt water return as adults, it must be assumed that the mortality of coho fry and trout in the Keogh river approached 100 per cent. 29

The Keogh river courses through very flat terrain, its valley within the spray area is not well differentiated, and although the forest cover is not dense it largely obscures the actual channel from aerial observation. The lower four miles of the watershed, lying in a cedar-hemlock swamp, were only partially sprayed, as a fisheries benefit, and this may account for the low initial mortality at station VI. It was apparently impossible however to avoid directly spraying the remainder of the system. The results of bottom sampling show a drastic reduction in aquatic insects at all three stations.

(3) Unnamed Creek-Station V Only the headwaters of this stream, several miles above the pen site were sprayed. There was no measurable initial mortality of either fry or fingerlings and the bottom samples did not show a severe reduction after spraying. Since this is a slow moving stream choked with debris in the upper reaches, it is considered possible that the DDT was removed by adsorption before it could reach the live pen.

(4) Coho Creel7-Station VII This stream is small, has a dense forest cover and would not have been easily avoided. It was sprayed at a dosage of one-half pound per acre, as was the adjoining Nimpkish area. A 31.8 per cent total mortality occurred the day after spraying but this did not increase in the following five days and there were very few distressed fish observed in the stream. This relatively low mortality is probably attributable to a combination of good forest cover and the lower dosage of insecticide.

(5) Nimpkish River-Station VIII The fry mortality at station VIII is listed in Table I as 95.9 per cent. The portion of the system sprayed however, between Nimpkish lake and the mouth of the lower Nimpkish river, is not utilized by coho and was presumably barren of juvenile salmon during the insect control program. Fry for the test pen were transported from Coho creek. Although it is assumed that there was no mortality in the Nimpkish river itself, the insecticide which caused a loss in the pen was carried into the estuary, killing large numbers of chum salmon fry and a few juvenile coho and spring salmon which had moved into the estuary from Johnston Straits. The Nimpkish is a large, wide river, easily seen from the air and is situated in an area that was treated with a one-half pound per acre dosage of DDT. The spray company agreed to fly the Nimpkish parallel to its course and to keep one swath width away from the river-bed. This stream should have been the most easily avoided in the spray area. The toxic concentration of insecticide must be attributed either to human error or to drift of spray. _

(6) Waukaas River-Station IX Station IX, sprayed on June 10, incurred a calculated DDT môrtality of 93.2 per cent. A heavy oil slick was observed on the 10th, a 35 per cent mortality was recorded by the 11th, on the 12th numerous dead coho fry and trout yearlings were observed in the stream and the pen mortality had risen to 89 per cent. On June 15 a few coho fry were observed alive in the stream but by October there were virtually none in evidence. The October bottom sample was almost barren, the total catch being one mayfly nymph and one annelid in 10 square feet. The Waukaas river lies in very flat terrain and has a medium forest cover. It would have been fairly difficult to avoid and the assessment results indicate that it received a full dosage of spray.

(7) Benson River-Station XI The Benson river, which flows into Alice lake on the Marble system, drains a narrow steep valley and the river channel is quite visible from the air. Prior to spraying, the pilots felt that this stream could be avoided. The assessment results and observations indicate however that the actual mortality was severe, being recorded at 72 per cent in the pen. On the afternoon of spray day, June 18, a number of dead coho fry, yearling trout and sculpins were seen at the mouth of the river on Alice lake, and it was noted that an extremely heavy oil slick extended one-half mile into the lake. An oil slick was observed also on three lakes in the upper Benson watershed and a trout kill was reported in Victoria lake, which adjoins Alice lake. There were no fry observed in the Benson river during the October survey and the bottom samples were extremely poor in aquatic insects.

(8) Colonial River-Station XII The corrected mortality at the station XII pen, which was sprayed on June 14, is listed as 94.9 per cent. The stream was checked on June 19, five days after spraying. There were no dead fish of any kind in evidence but a reduction of coho fry was reported, based on observation only. In October, coho fry were fairly numerous and the yield of the bottom sample, although lower than that of control streams, was not poor.

The actual mortality of coho fry in Colonial creek was probably light. Heavy rain occurred on June 10 and June 16, increasing the volume and velocity of the stream flow. The pen was situated in an unprotected site and much of the mortality was probably caused by excessive velocity. The valley of Colonial creek is not steep but the river channel is quite visible from the air. In addition, the south-east side of the valley was not sprayed. These two factors undoubtedly contributed to the success in avoiding severe mortality in this stream.

(9) Ingersoll River-Station XIII There was no mortality, attributable to the spraying, recorded at station XIII. The river is situated in a well-defined, unlogged watershed and is easily located from the air. Observers report that the aircraft did not appear to spray within one-half mile of the river itself and also that no oil slick was observed at any time. The area was sprayed on June 19, there was a heavy rainfall on the 21st and 22nd and the pen was removed June 23. The number of aquatic insects in the bottom sample, taken in October was comparable to that of the control streams and there appeared to be no depletion of coho fry. 31

(10) KlaslZish River-Station XIV The DDT mortality of coho fry in the Klaskish river pen has been calculated at 83.8 per cent. On June 18, three days after spraying, several hundred coho fry and few cutthroat trout, measuring up to eleven inches long were found dead at the mouth of the river. A drastic reduction in the number of coho fry present in the stream was also reported at that time. In late October, a bottom sample was taken and several seine hauls were made. There were no coho fry either caught or observed but the bottom sample, although lower in aquatic insects than those of the control streams, was much better than the samples taken on the Benson, Waukaas or Keogh rivers which suffered similar fish mortalities. Only the lower two and one-haif miles of this stream, which includes most of the spawning and rearing area, were sprayed. The insect populations were probably being repopulated from the upper areas. The Klaskish river is situated in a mountainous area and is clearly visible from the air. Observers report that on at least one occasion, an aircraft crossed this stream emitting spray but in general the pilots did attempt to avoid directly spraying the river channel. The observers reported also, howe:ver, that the spray drifted very badly. The fish mortality in the Klaskish river was apparently severe, that of the coho fry exceeding 80 per cent.

Bottom Samples A series of nine bottom samples were taken at stations I to VI throughout the period from May 27 to June 24 and two further samples were taken on July 25 and 30. Between October 8 and 21 samples were taken at these same stations, with the exception of number V, and in the Waukaas, Colony, Benson, Colonial, Ingersoll and Klaskish systems. Seven streams outside the spray area but in the same general region were also sampled for bottom organisms during this period. These were White creek, a tributary of the Quatse river, the Koprino river and Colony creek flowing into Quatsino Sound, the Spruce river at Holberg and the San Josef and Fisherman river. on the extreme northern tip of Vancouver Island. The individual bottom samples taken from May to July were generally of one square foot in size, the October samples inside the spray area were ten square feet and those taken outside were five square feet. The organisms of each bottom sample were enumerated, classified, dried and weighed. In general, mayfly larvae were the most numerous of the insect life, followed by stonefly and then caddis-fly larvae and various species of diptera and coleoptera. A species of oligochaete was present in most of the samples and at the locations where aquatic insect life was severely reduced, was sufficiently numerous to mask the loss of other forms on the basis of total number or total weight of organisms per unit area. Since these oligochaetes are not utilized for food by fish they were removed from the sample. Forms such as turbellaria, hirudinea and gastropoda were not common. With one exception, the loss of bottom organisms in the individual streams parallels closely the loss of fish. Table II lists thé weights and number of organisms per square foot of stream bottom for both spray and control streams from May to October. The weight data is shown graphically in Figure 3. Ga q PRE -SPRAY N

® POST- SPRAY (June-July)

M POST - SPRAY (October)

(Control)

I ID IY Y M Ig %I XŒ lm âIY STATION NUMBER

Ficuxe 3. The average weight of bottom organisms at several stations before and after spraying. 33

Table II The weight and number of organisms per square foot of stream bottom prior to spraying, immediately after and in October, Oligochaetes were excluded.

Stream Station Prior to After Spraying After Spraying Spraying (June-luly) (October)

STATION I ...... Number 23.5 5.25 6.7 (Cluxewe River) ...... Weight (Grams) 0.0223 0.0124 0.0024 it STATION 1 1 ...... 42.0 0.78 2.2 (Keogh River) ...... 0.0285 0.0038 0.0021 it STATION III ...... 26.9 5.21 5.9 (Keogh River) ...... 0.0226 0.0029 0.0018 « STATION IV ...... 24.5 12.6 13.0 (Quatse R. Control) ...... 0.0287 0.0212 0.0295 « STATION V ...... 21.7 20.0 (Unnamed Cr.)...... 0.0144 0.0103

U STATION VI ...... 37.3 9.2 1.2 (Keogh River) ...... 0.0500 0.0147 0.0004 is STATION IX ...... 0.1 (Waukaas River)...... 0.00002 « STATION XI ...... 0.8 (Benson River) ...... 0.0010 is STATION XII ...... 11.4 (Colonial) ...... 0.0058

cc STATION XIII ...... 17.2 (Ingersoll River)...... 0.0065

STATION XIV ...... it 5.9 (Klaslcish River) ...... 0.0040 cc Average of 6 ...... 19.1 Outside Streams ...... 0153

Streams in which there was a severe fish mortality such as the Keogh, Waukaas and Benson rivers, demonstrated a drastic reduction in aquatic insect life. The number and weight of organisms in others such as the Ingersoll, Colonial and Cluxewe rivers, which did not suffer a large fish mortality, were within the range of the control stream samples taken in October. The single exception to this parallel was the Klaskish river. Although results from the live pen on this stream indicate a mortality to coho fry of over 80 per cent, and observations suggest that this measurement is minimal, the bottom sample does not show a paucity of insect life. Only the lower two and one-half miles of the Klaskish, which includes most of the available spawning area, were sprayed. By October, when the sampling was carried out, the insect life could have been repopulated from the upper areas. As already stated, the most numerous forms priôr to spraying were mayfly, stonefly and caddis•9y larvae in that order. Table III compares the average number of each of these forms at stations I, II, III, V and VI; prior to spraying in June, and after spraying in June, July and October. The mayfly larvae, although suffering the greatest reduction in numbers, were still the most numerous after spraying. Caddis- fly larvae by October were non-existent in the samples.

Table III Comparison of the average number of Plecoptera, (Stone-flies), Ephemeridia (May-flies) and Trichoptera (Caddis-flies) per square foot of stream bottom before spraying, immediately after and in October at stations sprayed in the intensive study area.

Total Area - Station Plecoptera Ephemeridia Tricoptera Sampled in Sq. ft.

Prior to Spraying ...... I, II, III 2.5 20.0 4.0 22 (May-June) ...... V, VI After Spraying ...... II 0.7 2.4 1.1 46 (June-July) ...... : ...... After spraying ...... I, II, III, 1.2 1.7 0 50 (October) ...... VI

Table II compares the number and weight of bottom organisms for the same three periods as above for each station and stream sampled. In general, the proportion of larger forms in the sprayed stream samples was considerably reduced after spraying. At stations I, II and III, the number of organisms per unit area showed an increase over the June-July samples but the average weight declined. Both the number and the weight of organisms were lower in the October sample at Station VI on the Keogh river, but the weight per unit area was disproportionately low. In summary, the major fish food organisms of the Keogh, Waukaas and Benson rivers have been reduced to drastic proportions and will be of insignificant food value for at least the 1957 season. The abundance of major organisms in the other streams sampled varies, but the overall food production should not be greatly affected.

Bio-Assays In conjunction with the assessment programme, a series of bio-assay tests were conducted at the Nanaimo Biological Station of the Fisheries Research Board of Canada, to determine the toxicity of the spray formulation and its component parts. The results of this bio-assay programme have been reported by Alderdice and Worthington (see page 41, this issue). On the basis of the bio-assay tests, Alderdice considered the "safe" concentration of spray formulation to be below 0.05 parts per.million and that for the emulsifier, Atlox 2082A, to be 2.1 parts per million. Coho salmon underyearlings were used as subjects for the tolerance studies.

Analysis of Water Samples Water samples were taken at the three stations (II, III, and VI) on the Keogh river and at station IX on the Waukaas river. It must be emphasized that these were 35

two of the streams most affected by spraying. The samples, taken between June 10 and 22, arrived at the Biological Station on July 12. They were placed in storage at 00 to 40 C. on arrival and were analyzed for DDT content between October 2 and 7. It would appear that the pH of the water samples had risen on standing and the possibility exists that some degradation of DDT in the samples may have resulted. Results of the analysis are listed in Table IV.

Table IV Results of analysis for DDT content of water samples taken in the area sprayed for budworm control, Vancouver Island, June 1957.

Station Date Time Date DDT. P=.01 pH of Sampled Analyzed ppm. limits. ppm. sample

(Oct. 24, 1957) II ...... 10 June 0800 7 Oct. 0.13 0.09 -0.19 9.6 10 June 1930 3 Oct. 0.22 0.16 -0.32 9.7 (sprayed 11 June 0900 3 Oct. '0.20 0.14 -0.29 9.8 June 10) 12 June 0930 3 Oct. 0.03 0.019-0.044 9.85 13 June 1100 2 Oct. 0.22 0.16 --0 .32 9.9 14 June 1000 4 Oct 0.18 0.13 -0.26 9.6 18 June 0900 7 Oct. <0.01 - 9.7 20 June 0930 2 Oct. 0.40 0.28 -0.58 9.65

III ...... 10 June 1000 7 Oct. <0.01 - 9.7 19 June 1000 7 Oct. <0.01 - 9.85 (sprayed 19 June 1830 2 Oct. 0.37 0.26 -0.52 9.75 June 19) 20 June 0730 2 Oct. <0.01 - 9.9 21 June 0730 7 Oct. 0.03 0.019-0.044 9.75 22 June 0730 3 Oct. 0.10 0.07 -0.15 9.8 22 June 0800 2 Oct. 0.15 0.11 -0.22 9.5

VI ...... 10 June 1330 3 Oct. 0.07 0.048-0.11 9.75 11 June 1100 4 Oct. 0.28 0.19 --0.39 9.75 (sprayed 12 June 1300 2 Oct. 0.21 0.14 -0.30 9.75 June 10 and 12) 18 June 1200 7 Oct. 0.01 - 9.65

IX ...... 10 June 1200 4 Oct. 0.06 0.042-0.090 9.6 11 June 0930 4 Oct. 0.05 0.033--0.074 9.65 (sprayed 12 June 1600 3 Oct. 0.07 0.048-0.11 9.8 June 10) 14 June 1130 4 Oct. 0.02 0.013-0.031 9.1

The Waukass river was sprayed on June 10. On that date, the DDT content, as measured, was 0.06 p.p.m., and three days later the concentration was found to be 0.07 p.p.m. However on June 14 the concentration of DDT had dropped to 0.02 p.p.m., less than the safe level of 0.05 p.p.m. (Table IV). It can therefore be stated that a toxic concentration of the insecticide was present in the Waukass river for at least three days. As outlined, previously, on the Keogh river, station VI was situated near the mouth, station II eight miles upstream and station III six miles above that. Most of the watershed below the upper station was sprayed on June 10 although a small part of the lower area was treated on June 12. The uppbr watershed, which included station III, was sprayed on June 19. A water sample taken at the middle station 12 to 14 hours after spraying on June 10, on analysis, showed a concentration of 0.22 p.p.m. of DDT while that of a sample taken four days later was measured at 0.18 p.p.m. By June 18, eight days after spraying the concentration of DDT had dropped below the measurable limit of 0.01 p.p.m. but in a sample taken two days after the June 19 spraying of the upper area, the concentration had risen to 0.40 p.p.m. The middle portion of the Keogh river was subjected therefore to at least five days of toxic DDT levels. A water sample taken at Station III, sprayed on June 19, had a DDT content of 0.37 p•p.m. on spray day. Four days later the concentration had only dropped to 0.15 p.p.m., very much above the safe level. Water samples were taken at the lower station (VI) on June 10, 11, 12 and 18. In the first two days after spraying, the concentration was very high but by June 18, had dropped below the 0.01 p.p.m. limit. Since sampling was discontinued at this station on June 18, the effect of the upper watershed spraying of June 19 was not measured. This was the station at which there was no measurable initial fish mortality but at which there were no juvenile salmonids present in October.

Discussion As stated earlier, the hazards of aerial spraying with DDT to fish and fish-food populations were fully realized by the fisheries groups prior to spraying. Although it was also realized that the preventative measures agreed upon would not eliminate entirely the probability of severe fish mortality, they did constitute an improvement over current practices. The success of avoiding mortality was therefore dependent upon the ability of the pilots to effect those measures. The actual spraying was kept under close observation by fisheries personnel. The air crews were well experienced in this type of spraying and all those concerned with the aerial control program, including the pilots, expressed repeatedly the desire to avoid damage to the fish populations within the area. The fisheries observers agreed that in general the pilots made a sincere attempt to carry out the preventative measures agreed upon. Presumably, the equipment and personnel used in the insect control programme were of the highest calibre, all practical preventative measures (short of non-spraying) were followed and a sincere effort made to implement those measures yet the mortality in at least four major streams approached 100 per cent. The meteorological conditions during the spray period, although possibly adverse for spraying, were not unusual for the region at that time of year. There was a large variation in the effect of the DDT application on the fish population. The mortality of coho fry ranged all the way from zero on the Ingersoll to almost complete annihilation on the Keogh and Waukaas rivers. The reduction in aquatic insects paralleled quite closely the loss of fish. Although the effects of spraying varied with each stream, the following points might be noted: (a) Fish mortality was high in large streams flowing through flat terrain with a fairly dense forest cover (e.g. Keogh and Waukaas rivers). The probable reason is that they are not clearly visible from the air. 37

(b) Fish mortality was high in large streams flowing through steep-walled valleys (e.g. Klaskish and Benson rivers). Watersheds of this type are sprayed on contours and in the particularly rugged terrain with its attendant unpredictable air currents, the control over spray deposition must be very slight. (c) Fish mortality was low in streams situated in well differentiated but not particularly steep-walled valleys (e.g. Ingersoll and Colonial rivers). These streams are visible from the air and there is apparently better control of spray deposition. (d) Fish mortality was relatively low in a small stream with a dense forest canopy (e.g. Coho creek). This stream was sprayed at a dosage of one-half pound of DDT per acre. In view of the heavy mortality of chum salmon fry in the Nimpkish river, which was also sprayed at the same dosage, it must be concluded that the reduction in concentration was not the significant factor in the low mortality in Coho creek. (e) There was no measurable initial fish mortality in a small stream which was sprayed in the upper area only, several miles above the live pen (e.g. Unnamed creek). It is considered probable that the concentration of DDT was reduced below toxic levels by adsorption.

Loss of Salmon Except for the loss of many thousands of chum fry which were killed in the estuary of the Nimpkish river, the damage to salmon stocks was confined to the mortality of coho fry. There are 18 salmon streams within the spray area, excluding the lower Nimpkish river which does not serve as a spawning or rearing area for coho, and in ten of these the 1956 coho escapement was estimated at 500 or more. The effects of spraying were measured in nine of those ten major streams as shown in Table I. In four of these, the Keogh, Waukaas, Klaskish and Benson rivers the mortality of coho fry approached 100 per cent and in a fifth, Coho creek, the loss was calculated at 30 per cent. In the four others that were sampled, the Ingersoll, Colonial, Cluxewe, and Unnamed creeks, there was not a large mortality. Mills creek, south of Port McNeill, was the only major stream where mortality was not assessed. The damage to salmon stocks within the spray area was over-shadowed by the loss of coho fry in the Keogh river, where the progeny of an estimated escapement of 40,000 adult coho were almost completely annihilated. The escapement to the other three major streams affected by spraying totalled between 2,000 and 4,000. The coho salmon has a predominant three-year life cycle and very little overlapping from other year classes can be expected. Even assuming that there was not a complete loss of fry in the four grossly affected streams, and assuming also that there will be some smolt production, judicious management will not restore the population for many cycles. The economic loss will then be accumulative until the escapement is restored once again to the 1956 level. In order to alleviate this loss, the most obvious assistance to management would be artificial restoration. The only immediate measure would be the release of underyearlings, preferably from neighbouring streams. The problem of successfully trans- porting significant numbers to the streams in this isolated region, if the underyearlings were available, would be extremely difficult. In addition, it is doubtful if the streams in their present unproductive state could support a significant transplant. The best chance for success of artificial restoration may be the introduction in 1959 of large numbers of eggs. These could be flown to the Port Hardy airport and distributed to the respective streams.

Loss of Trout Rainbow, cutthroat, and steelhead trout are indigenous to all streams in the region. It was not possible to estimate the number of trout within the spray area and knowledge on the size of the steelhead trout escapement is lacking. Unlike Pacific salnion, the steelhead and both species of trout spawn in the spring and by spray time in mid-June the spawn had probably reached the alevin stage. It is highly probable that the alevins are susceptible to DDT and in fact, a paucity of fry of the year were observed in October within the spray area. Seining was conducted in two control streams, the Koprino and Quatse, and trout fry were present in normal numbers. In general, steelhead trout spawn in the spring, the fry emerge in the early summer, spend the remainder of that year and the whole of the following year in fresh water before going to sea in the late spring of the third year. In 1957, the seaward migration of smolts was well underway by spray date and only the latter portion would have been affected by DDT. Results from the holding pens indicate that the initial mortality to steelhead smolts was relatively light as compared to that of coho fry. The loss of migratory steelhead smolts, consequently, may not have been severe. The real damage to the steelhead populations would have been the loss of underyearlings and very probably the spawn of the year, in which case two consecutive year classes would have been affected. This loss was particularly noticeable in the Keogh river, where in October, there was a near complete absence of all juvenile salmonids. In general, the relative mortality of trout in the affected streams could be expected to parallel the loss of coho fry. If this assumption is correct, of the major streams, the greatest losses would have occurred in the Keogh, Benson, Klaskish and Waukaas rivers. Of these four, only the Keogh is utilized at present to any degree by anglers, but the other three do have a potential. The resident trout of the Keogh river may repopulate quickly from several small lakes on the system but regeneration of the steelhead population will be a slow process. The trout population in the Benson river will likely be rehabilitated from the resident trout populations of Alice Lake. In October, angling was good in the lower portion of the stream. The Klaskish river is situated on the west coast of Vancouver Island and because of the isolated region, will not be an important angling stream for many years. Only the lower two and one-half miles of this stream was affected by spraying and repopulation should proceed rapidly from the upper areas. The Waukaas river is not a major angling stream, but it is considered that the resident trout and fresh water stages of steelhead trout were eliminated by spraying. A few ta.kes received marginal doses of insecticide and although trout mortality was reported in the two largest, Alice and Victoria, it is considered that generally the losses were relatively small and will be quickly replaced by natural propagation. 39

In summary, the loss in number of trout by spraying was undoubtedly large. This district is however isolated and under the relatively light angling pressure imposed upon them at present, the stocks of trout should repropagate to pre-spray levels quickly. The steelhead trout on the other hand will take a number of years to overcome the effects of spraying which undoubtedly affected two consecutive year classes.

Reduction of Aqudtic Insects The reduction in aquatic organisms within the spray area as measured by bottom sampling, followed the same pattern reported by Ide (1956). The major loss was confined to aquatic insect larvae, particularly the larger forms of stoneflies, mayflies and caddis flies. Mayfly larvae were the most numerous form both before and after spraying but also suffered the great percentage reduction. Caddis-fly larvae by October were almost absent in the samples taken from the four major streams affected. Ide (1956) reports that in a tributary of the Miramichi, which was included in the extensive Maritimes spraying of 1954, the emergence in the year of spraying consisted of large numbers of minute chironomids and very few of the larger insects. In the year following spraying the bulk of insects emerging was less than half that of the control stream and that was made up predominantly of small chironomids. There was a lack of caddis flies. In the second year after spraying, the bulk of the emergence from the spray stream was still considerably lower than that of the control, and minute chironomids continued to dominate the emergence of the former. There were fewer species in the spray stream, stoneflies were still present in only insignificant numbers and large forms of stoneflies and mayflies were absent from the samples. Kerswill (1957) found on the Miramichi that the normal diet of underyearling Atlantic salmon consists largely of small insects, but that older juveniles feed mainly on larger forms and do not appear to utilize the small forms even when the latter are abundant. On the basis of these findings, the productivity of streams that are destitute of aquatic life at present will remain low for at least two seasons.

Summary During mid-June of 1957, an aerial spraying programme was conducted on 155,000 acres of timberland on the northern portion of Vancouver Island, in an attempt to control an outbreak of black-headed budworm. The formulation used was one pound of DDT in a solvent with an emulsifier added and blended and one U.S. gallon with diesel oil. This was applied at the rate of one U.S. gallon per acre. The damage to the fish and fish-food populations was assessed on the major streams and on four of these, was found severe. The fish mortality was confined generally to coho fry, trout, steelhead yearlings and possibly alevins of both trout and steelhead. In the four major streams affected by spraying, the progeny of an estimated 1956 escapement of 43,000 coho adults and the juvenile stages'of several thousand steelhead and trout was almost eliminated. The reduction of aquatic insects parallels the loss of coho fry and the productivity of several streams is not expected to return to adequate proportions for at least two years. A series of bio-assays was conducted at the Nanaimo Biological Station of the Fisheries Research Board of Canada. The tests indicate that a safe concentration of the formulation used in this insect control programme is below 0.05 parts per million. Analysis of water samples taken in the field showed that toxic concentrations of DDT existed at four test stations for more than three days after spraying.

Literature Cited

ALDERDICE, D. F. and M. E. WORTHINGTON 1958, Toxicity of a DDT forest spray to young salmon. Canadian Fish Culturist, this issue, pp. 41-48. IDE, F. P.1956, Effects of forest spraying with DDT on aquatic insects of salmon streams. Trans. Amer. Fish. Soc., 86, pp. 208-219 KERSWILL, C. J. 1957, Investigation and managnunt of Atlantic salmon in 1956, Trade News, No. 12 pp. 5-15 Toxicity of a DDT Forest Spray to Young Salmon by

D. F. Alderdice and M. E. Worthington

Fisheries Research Board of Canada, Biological Station, 7lanaimo, B. C.

In June, 1957 an aerial DDT spray programme was carried out to control black- headed budworm in an area of northern Vancouver Island, B.C. This report sum- marizes a series of laboratory tests conducted to determine the tolerance of a representative species of Pacific salmon to the DDT formulation used in the spray programme. The study provides information relating to an extensive field survey conducted to assess the effects of spray deposition on aquatic fauna within the sprayed area (Crouter and Vernon, 1959, see page 23, this issue). Method

Following consultation with members of the Department of Fisheries and B.C. Game Commission, the following programme was undertaken: (a) Examination of tolerance to Atlox 2082A, the spray formulation emulsifier. (b) Examination of tolerance to the spray formulation used in the aerial spray programme. (c) Comparison of the potency of the aerial spray formulation with a DDT- acetone standard. Accordingly, the following materials were used in the tolerance tests: (a) Emulsifier-Atlox 2082A, a preparation of "alkyl aryl sulphonate" and "polyoxyethylene sorbitan esters of mixed fatty and resin acids" manufactured by the Atlas Powder Company. (b) Aerial Spray Formulation-made up in the following manner: 25 g. 100% Tech. "Montrose" DDT 7.14 g. Base Oil 3.6 g. Atlox 2082A One volume of the above mixture is diluted with 1.2 volumes of medium grade diesel oil. The "base oil" used was "Standard Base Oil (Wood Treating)" manufactured by the Standard Oil Company. "Medium grade diesel oil" was obtained from the Imperial Oil Company of Canada. DDT was obtained from the Stauffer Chemical Company, Portland, Oregon. (c) DDT-Acetone Standard-Maintaining the same proportions of DDT and Atlox 2082A as in the aerial spray formulation, acetone was substituted for the base oil and medium grade diesel oil. ,

41 42

Estimation of DDT content of solutions used in the tolerance tests and of the water samples collected in the spray area was carried out by a modification of the method suggested by Amsden and Walbridge (1954). A standard curve was prepared by analyzing for known amounts of DDT in acetone solution. No significant difference was obtained in a comparison of results from these standards with those obtained using aerial spray formulation added to water with subsequent separation of DDT with isopropyl ether. The resulting standard curve relating DDT concentration to spectro- photometric absorbance was used for estimating DDT levels in the succeeding tests. In all cases absorbance corrections were made with reagent blanks run concurrently. The method provided measurement of DDT down to 0.01 p.p.m. after extraction of the DDT from water. Colorimetric determinations were made using a Beckman DU Spectrophotometer at a wave length of 4700 A° and 0.03 mm. slit width. Coho salmon (Oncorhynchus kisutch) underyearlings were considered as an appropriate species for determination of tolerance to DDT. The tests conducted with Atlox 2082A were carried out in August, 1957, and the coho underyearlings used averaged 6.89 cm. in length and 4.08 g. in weight. Test fish were exposed to various concentrations of Atlox 2082A at a ratio of 0.88 g. of fish per litre of solution. The tests conducted with the aerial spray solution and DDT-acetone standard were carried out in October, 1957. The coho underyearlings used averaged 8.81 cm. in length and 8.20 g. in weight. Test fish were exposed to various concentrations of the two materials at a ratio of 1.8 g. of fish per litre of solution. Fish used in the experiments were acclimated to and tested at 10°C. One-half (23 litres) of the total volume of each test container was exchanged daily to minimize the possible diminution in DDT content by loss to the test fish. Daily determinations of DDT content were made by withdrawing one litre of each test solution and extracting the DDT with isopropyl ether. Such samples were taken at the beginning of each experiment, and immediately before each 23 litre solution exchange. Results Toxicity of Atlox 2082A to coho salmon underyearlings Samples of 10 fish were exposed to each of eight concentration levels of Atlox 2082A from 500 p.p.m. to 10 p.p.m., along with a control. Times to 50 per cent

Table I Time to 50 per cent sample mortality (ET5o) for samples of 10 coho underyearlings exposed to various concentrations of Atlox 2082A in fresh water.

Concentration of Atlox 2082A, p.p.m. ET60, minutes

500 33 100 98 70 290 50 380 40 680 30 1,125 20 2,450 10 > 12,517 Control No mortality 43

10 100 1000 Atlox 2082 A, pp.m.

Ficuna 1-Times to 50 per cent mortality in samples of 10 underyearling coho exposed to various concentrations of Atlox,2082A. The arrow in the upper corner refers to a test at 10 p.p.m. which was discontinued at 12,517 minutes. mortality in each concentration were calculated from the distribution of mortality tunes. Results of the tests are listed in Table I and illustrated in Figure 1. Analysis of the transformed data provides an equation for the line from 20 to 70 p.p.m. From the equation the 48 hr. TI,m ^median tolerance limit) (see Doudoroff, et at 1951) may be inferred and is estimated as 20.7 p.p.m. of Atlox 2082A. A safe level of Atlox 2082A, in terms of the experimental procedure, is estimated as 0.1 x 48 hr. TL.m or 2.1 p.p.m. Atlox 2082A. On the basis of the aerial spray formulation a unit of the spray in water containing 1 p.p.m. of DDT would contain about 0.15 p.p.m. of Atlox 2082A. Therefore, the concentration of DDT in a unit of water containing 2.1 p.p.m. Atlox 2082A would be approximately 14.0 p.p.m. DDT.

Toxicity of Aerial Spray Formulation to Coho Salmon Underyearlings Samples of 10 fish were exposed to each of six concentration levels of the aerial spray formulation in water. The formulation does not form a stable emulsion in water and partition of the oil and aqueous layers occurred very rapidly. It was assumed for this reason that only a limited amount of the theoretical doses of DDT applied would be available throughout the water mass. In the hikher concentrations tested a film of oil was visible on the water surface of the test containers. Subsequent tests also indicated a considerable amount of the DDT in the test tanks was being adsorbed upon sediment in the freshwater supply and precipitated with that sediment to the bottom of the tanks. The latter fact has been well documented in the literature (see Berck, 1953). For these reasons the remaining concentrations established for tolerance tests are reported both as theoretical doses (level of DDT concentration by volume dilution) and as actual doses (level of DDT by chemical analysis of water taken midway between the surface and bottom of the test containers). Results of the tests on the aerial spray formulation are listed in Table U. Probit response curves for the concentrations examined are illustrated in Figure 2. The ET60 values derived from these curves are illustrated in Figure 3. The median resistance times describe a most unusual response distribution. Maximum résponse in terms of time to death occurs at a dilution of the spray formulation providing a theoretical dosage of 0.5 p.p.m. (actual dosage of 0.05 p.p.m.). Test fish were less susceptible to the solutions at dosages both less and greater than 0.05 p.p.m. DDT (actual dosage).

♦ ♦

'à-3.0 •-1.0 0.7 0-0.5 n - 0.3 O - 0.1

5000 10000 RESISTANCE TIME (Min.) FIGURE 2-PRobit response curves illustrating the rates of mortality of coho underyearlings exposed to various concentrations of the aerial spray formulation. The curves are identified as to the theoretical concentrations of DDT applied, in p.p.m. i 45

Table II Time to 50 per cent mortality (ET5o) for samples of 10 coho salmon underyearlings exposed to various concentrations of the aerial spray formulation.

Dilutions of aerial spray formulation providing DDT Average concentration at the following theoretieal of DDT by chemical ETso, minutes concentrations, p.p.m. analysis, p.p.m.

3.0 0.36 > 12,492 1.0 0.19 4,400 0.7 0.11 3,600 0.5 0.05 2,950 0.3 0.10 2,950 0.1 0.07 > 14,459 0.03 0.02 > 10,031 Control Control Nô mortality

T /T i ^ ... m n_^ 1000' E %* -0- O

DDT, theoretical

DDT, by analysis

L I I I I I 1 11 1 1 1 1 111Lt u 0.01 0.1 1.0 p.p.m. DOT in aerial spray formulation

FIcunE 3-Times to 50 per cent mortality in samples of 10 underyearling coho exposed to various concentrations of the aerial spray formulation. The arrows refer to tests in which 50 per cent mortality had not occurred at the times when the tests were discontinued. A further observation made in the 3.0 p.p.m. theoretical concentration (0.36 p.p.m. by analysis) supports field evidence of the Department of Fisheries (private communication). After 11,000 minutes of exposure almost all fish still living showed symptoms of blindness. Quite evident was the opacity of the lens of one or both eyes of these remaining fish.

Toxicity of DDT-Acetone Standard Samples of 10 fish were exposed to each of six concentrations of a DDT-acetone standard. Although a highly stable emulsion was produced by the mixture of these materials, there was a tendency for the DDT to crystallize out onto the surface of the water solution, apparently on evaporation of the acetone carrier. Again it was sus- pected that. DDT would be adsorbed onto particulate matter in the water, with a resulting diminution in the amount of DDT available throughout the water mass. Results of this series of tests are listed in Table III. The distribution of median resistance times is illustrated in Figure 4.

Table III Time to 50 per cent mortality (ETm) for samples of coho salmon underyearlings exposed to various concentrations of the DDT-acetone standard.

Dilutions of the reference solution providing DDT at Average concentration the following theoretical of DDT by chemical concentrations p.p.m. analysis, p.p.m. ET5o, minutes

3.0 0.31 850 1.0 0.29 730 0.7 0.06 850 0.5 0.05 1,100 0.3 0.09 1,150 0.1 0.08 1,750 0.07 0.05 > 10,075 Control Control No mortality

In this series nearly all test fish died within the first 48 hours, consequently few analyses were possible for actual DDT content of each test solution. Some of the error in the response distribution for this series is undoubtedly connected with inability to sample at shorter intervals. A further source of error would be associated with variations in the amount of particulate matter in the test solutions. Analysis of the transformed data indicates considerable variation in the distribution of resistance times; nevertheless, the trend in median resistance times is considered to be linear (Figure 3). Abrupt changes in slope within the probit response curves from 0.1 to 0.7 p.p.m. DDT theoretical dosage suggested that the response of the fish to the solutions might have been changing with evaporation of the acetone carrier. Interpretation of Results A limiting concentration of Atlox 2082A is estimated to be approximately 2.1 p.p.m. and, proportionately, this amount would be found at a DDT level of 14.0 p.p.m. The action of Atlox 2082A as a toxic agent would probably be found only at 47

.► O t^ F- W

0.01 0.1 1.0 p.p.m. DDT in acetone standard

FtcuttE 4-Times to 50 per cent mortality in sampks of 10 undc+yearling coho exposed to various concentrations of the DD`I'-acetone standard. For an explanation of the arrows, set Fig. 3. extremely high concentrations of spray formulation. It is doubtful if such concentrations would be found under spray conditions. No valid statistical comparison can be made between the response distributions for the aerial spray formulation and DDT-acetone standard as that of the former is not linear. The bizarre distribution of resistance times for the aerial spray formulation suggests that the physical aspects of contact with the spray formulation or its components are dissimilar within the range of concentrations examined. In terms of DDT available in the water mass, the aerial spray caused mortality at and above 0.05 p.p.m. whereas 0.02 p.p.m. available DDT did not affect the test fish within the experimental period. In the case of the DDT-acetone standard, the lowest concentration producing mortality was again 0.05 p.p.m. available DDT. However, as the 0.07 p.p.m. test (0.05 p.p.m. available DDT) did not cause mortality within the experimerltal period, whereas the 0.5 p.p.m. (0.05 p.p.m. available DDT) did, factors may operate to determine toxicity other than that of the final amount of DDT available. Comparing the two response distributions from another aspect, the resistance times of the samples of young salmon exposed to the aerial spray are greater than those for the fish exposed to the same range of DDT concentrations in the DDT-acetone standard, even though this range ultimately was lethal in both cases. In summary, and in terms of the test conditions, the aerial spray formulation may be regarded as "safe" to young coho salmon below approximately 0.05 p.p.m. where the measure of safety is regarded as the absence of mortality within an exposure period of one week at 100C. An unusual response distribution was found for the aerial spray with resistance times increasing with increase in dosage over the higher portion of the concentration range examined. Associated with high concentration of spray, however, was development of blindness in the test fish. The series of tests tended to point up the complexity of the problem of interpretation of the action of associated variables. Besides the measure of DDT available to the test fish within the water mass, it is considered that the stability of the solution in such aspects as volatility, gravity stability, adsorptive potential of inclusions in the water supply and action of associated formulation components may have a considerable influence on the biological activity of the formulation used. Indeed, such factors may to a large extent be responsible for some of the diversity of tolerance estimates reported in the literature. A thorough study of the relation of such physical characteristics to biological effects on fishes would be a useful adjunct to such tolerance studies. It should also be recognized that the criterion of safety applied here for young salmon does not necessarily relate to the impact on the total biological community. The tests do indicate, however, the range of concentrations of DDT and the related exposure times which could be harmful in the field.

Literature Cited

AM3DEN, R. C. and D. J. WALBRIDGE. 1954. Simplified method of estimating DDT residues. J. Agric. Food Chem., 2(26): 1323-1324. BERCK, B. 1953. Minodetermination of DDT in river water and suspended solids, Analytical Chem., 25(8): 1253-1256. CROUTER, R. A. and E. H. VERNON. 1959. Effects of black-headed budworm control on salmon and trout in British Columbia. The Canadian Fish Culturist, this issue, pp. 23-40. DOUDOROFP, P., B. G. ANDER9ON, G. E. BURDICK, P. S. GALTSOFP, et al. 1951. Bio-assay methods for the evaluation of acute toxicity of industrial wastes to fish. Sew. and Ind. Wastes, 23(11): 1380-1397.

ISSUE TWENTY-FIVE OCTOBER - 1959

THE CANADIAN FISH CULTURIST

ETR^t 4RX I't--i-IERIES AND OCEAIV'S BIBLIOTHÈQUE PECHES ET OCFANg ï

Published at Ottawa by The Department of Fisheries of Canada