. /3/JL -' MM /iS NOAA TM NMFS ABFL-3 A UNITED STATES DEPARTMENT OF COMMERCE PUBLICATION NOAA Technical Memorandum NMFS ABFL-3 U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration National Marine Fisheries Service SEATTLE, WA NOVEMBER 1974 Salmon Fry Production in a Gravel Incubator Hatchery, Auke Creek, Alaska, 1971-72 JACK E. BAILEY and SIDNEY G. TAYLOR NOAA TECHNICAL MEMORANDA National Marine Fisheries Service, Auke Bay Fisheries Laboratory NOAA Technical Memoranda of the National Marine Fisheries Service Auke Bay Fisheries Laboratory deal with research conducted at that Laboratory, which is located at Auke Bay, Alaska. Copies of the NOAA Technical Memorandum NMFS ABFL are available from the Laboratory or from the National Technical Information Service, U.S. Department of Commerce, Sills Bldg., 5285 Port Royal Road, Springfield, VA 22151. NMFS ABFL-1. An improved incubator for salmonids and results of preliminary tests of its use. By Jack E. Bailey and William R. Heard. NMFS ABFL-2. A Guide to the Collection and Identification of Presmolt Pacific Salmon in Alaska with an Illustrated Key. By Milton B. Trautman. NMFS ABFL-3. Salmon Fry Production in a Gravel Incubator Hatchery, Auke Creek, Alaska, 1971-72. By Jack E. Bailey and Sidney G. Taylor. U.S. DEPARTMENT OF COMMERCE Frederick B. Dent, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator NATIONAL MARINE FISHERIES SERVICE Robert W. Schonmg, Director NOAA Technical Memorandum NMFS ABFL-3 Salmon Fry Production in a Gravel Incubator Hatchery, Auke Creek, Alaska, 1971-72 JACK E. BAILEY and SIDNEY G. TAYLOR ■riO WW*/*,, 'iff NT nt l>" >> a O u £ SEATTLE, WA o November 1974 a 4) c The National Marine Fisheries Service (NMFS) does not approve, rec- ommend or endorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales pro- motion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. CONTENTS Page Introduction 1 Building and water system 2 Experimental design 2 Water quality 2 Collection and pretreatment of eggs 4 Auke Creek eggs 4 Sashin Creek eggs 4 Incubator design and operation 4 Bams boxes 4 NMFS boxes 5 Tray incubator 6 Natural spawning in Auke Creek 6 Enumeration of fry 6 Collecting and processing samples of fry 7 Results 7 Water quality 7 Temperature 7 Oxygen 7 Sediment 8 Fry production and quality 8 Survival from egg to fry , . 8 Effects of density on survival from egg to fry 9 Effects of genetics and box design on survival from egg to fry 9 Effects of conditions in the incubators on size and stage of development of fry 9 Time of emergence of incubator-reared and wild fry 11 Discussion 11 Survival 12 Size and stage of development 12 Time of emergence 13 Acknowledgments 13 Literature cited 13 Figures 1. Site of incubation facility on Auke Creek, southeastern Alaska 2 2. Incubation facility housing gravel incubators and other equipment need- ed to rear salmon fry. Four incubators are shown at the lower left, and one incubator is being seeded with pink salmon eggs 3 3. Bams box used for incubating pink salmon eggs 4 4. Cross-section detail of Bams box 5 5. NMFS box used for incubating pink salmon eggs. 6 6. Cross-section detail of NMFS box 6 7. Temperatures of water in incubators and of surface water in Auke Creek, 21 August 1971 to 15 May 1972 8 8. Dissolved oxygen levels in the influent and effluents from four incubators at Auke Creek incubation facility, 1971-72 8 9. Lengths of preserved fry in relation to source of eggs and incubation en- vironment 11 10. Weights of preserved fry in relation to source of eggs and incubation en- vironment 11 11. Indices of development, KD, of preserved fry in relation to source of eggs and incubation environment 11 12. Daily cumulative percentage of emergence of pink salmon fry from each gravel incubator at Auke Creek incubation facility between 23 March and 14 May 1972 12 iii Tables Page 1. Density of eggs, water flow, and apparent velocity of NMFS boxes B and D and Bams boxes C and A 5 2. Fry production, flow rates, and oxygen consumption in four gravel incubators at Auke Creek, 1 February through 4 April 1972 8 3. Number of pink salmon eggs or fry at different life stages and survival be- tween stages for Auke Creek eggs incubated in gravel incubators. ... 8 4. Numbers of eyed pink salmon eggs and survival to emergent fry in indivi- dual incubators 9 5. Mean and variance of lengths, weights, and developmental indices of pink salmon fry from Auke Creek, gravel incubators (Bams and NMFS boxes), and tray incubators 10 6. Index trap catches of creek-incubated pink salmon fry migrating from Auke Creek, 11 April to 14 June 1972 12 IV Salmon Fry Production in a Gravel Incubator Hatchery, Auke Creek, Alaska, 1971-72 JACK E. BAILEY and SIDNEY G. TAYLOR' ABSTRACT Survival and physical characteristics of pink salmon fry, Oncorhynchus gorbuscha, incubated in two types of boxes, each box containing about 1 m1 of gravel, and a Heath in- cubator were compared with fry from natural spawning to evaluate the use of boxes to produce fry. The gravel incubators were seeded at densities of 74,200 to 198,000 eyed eggs/m3. Survival from eyed eggs to emergent fry ranged from 79 to 97% in artificial incubation, but the number of incubators tested was too small to define any relationships between survival and incubator type or egg density. With artificial incubation in gravel, survival from potential eggs in females to emergent fry was 69%, whereas with natural spawning and incubation in the creek, survival was about 12%. Fry emerged from gravel incubators about 3 days earlier than from the streambed. The gravel incubator fry were larger than tray fry but smaller than creek fry. The smaller size of the gravel incubator fry could not be explained entirely on the basis of early emergence. Further studies were recommended to determine whether the muskeg sediment that ac- cumulated in the incubators, the low oxygen level (57 to 69% saturation), or the substrate parti- cle size and composition inhibited growth of the embryos. INTRODUCTION Scientists of the National Marine Fisheries Service (NMFS) have conducted research in Alaska for over 10 yr to evaluate the environmental requirements of salmonid eggs, alevins, and fry in the natural streambed. To refine the studies of embryo ecology, the scientists at the Auke Bay Fisheries Laboratory developed a system of incubating salmonid eggs in boxes of gravel similar to the system developed by Bams (1970). Biological, chemical, and physical fac- tors such as egg density, oxygen levels, and water velocities could be controlled or monitored more precisely in the boxes than in a natural streambed. The boxes required a minimum of space and water in which substantial numbers of robust fry could be produced. In a laboratory test, 1-cubic-foot boxes seeded with 10,000 eggs of pink salmon, On- corhynchus gorbuscha, produced 8,300 viable fry per box — a fivefold to tenfold increase in survival over 'Northwest Fisheries Center Auke Bay Fisheries Laboratory, National Marine Fisheries Service, NOAA, Auke Bay, AK 99821. Taylor was employed by the Alaska Department of Fish and Game during most of the first year of this cooperative study with the National Marine Fisheries Service. The project also involves a third agency. Territorial Sportsmen, Inc. of Juneau, Alaska. The pilot gravel incubation facility was assembled on property owned by Territorial Sportsmen, Inc. that expected in natural redds (Bailey and Heard, 1973). Water flow in the boxes was about 2.0 liters/min. The fry from 1-cubic-foot boxes could not be distinguished from wild fry on the basis of size and ability to resist starvation in seawater. Similar in- cubators developed in British Columbia yielded a six- fold advantage, in numbers of returning adults, over natural reproduction of pink salmon (Bams, 1972). Thus, the concept of incubating salmon eggs in a carefully controlled gravel environment appears to have good potential as a method to increase fry production. Auke Creek was selected as the site for the test of the incubator boxes because it is near the Auke Bay Fisheries Laboratory and because an existing pipeline provides an adequate water supply with adequate protection against freezing from nearby Auke Lake (Fig. 1). Construction of the pilot production facility began in July 1971, and the incubation of pink salmon eggs began in the fall of that year. Although the in- cubator box concept may evolve into a technique for low-cost mass production of salmon fry in remote areas of Alaska, some preliminary field experiments (Bailey and Heard, 1973) have ended in catastrophic losses of eggs and alevins. These losses were due prin- cipally to mechanical difficulties in maintaining a continuous flow of water at remote locations during severe cold weather. We felt that the biological feasibility of the incubator concept could be evaluated Figure 1. — Site of incubation facility on Auke Creek, southeast- ern Alaska, where the National Marine Fisheries Service, NOAA, and Alaska Department of Fish and Game are experi- menting in a cooperative effort to rear salmon fry in gravel incubators. on a reliable water supply such as was available at the Auke Creek site. This report of the first 12 mo of progress (July 1971 through June 1972) describes the Auke Creek incuba- tion facility, the techniques used in collection of eggs and operation of the gravel and tray incubators, the water quality in the incubators, and the differences between incubator fry and creek fry. The immediate objectives for the first year of operation were (1) to use boxes of gravel (gravel incubators) to produce a sub- stantial number of pink salmon fry, (2) to compare survival and physical characteristics of fry from the gravel incubators with fry from Auke Creek and with fry produced on flat screened trays (tray incubators), and (3) to supplement the natural production of fry in Auke Creek by releasing fry of Auke Creek parentage from the gravel incubators. BUILDING AND WATER SYSTEM A 7.3- by 13.4-rn (24- by 44-foot) heated building (Fig. 2) provided space for gravel incubators con- taining at least 1 million eggs, four stacks of tray in- cubators, a 567-liter/min water filter and ultraviolet purifier system, equipment for fry censusing and sampling, and a storage room. The building was built on the bank of Auke Creek near a fish-counting weir (Fig.l) where eggs could be collected from returning adult salmon. Water was supplied from Auke Lake through a buried 10.2-cm (4-inch) polyvinyl chloride (PVC) pipe connected to a 35.6-cm (14-inch) wood stave pipeline; the intake was about 7.6 m beneath the surface and 1.5 m off the bottom of Auke Lake. The water was dis- charged into an elevated 1,000-liter fiber glass head tank inside the building, which ensured a constant 3.6 m of hydraulic head above the floor of the building. The head tank could be bypassed to supply untreated water to the incubators; the 10.2-cm bypass had about 7.9 m of hydraulic head at floor level. Because the filter-purifier system was not installed until February 1972, the bypass was not used for the 1971-72 incuba- tion period. One advantage of the Auke Lake water source was the relatively small risk of loss of flow because of freezing. A disadvantage of the lake water was the difference in temperature compared with Auke Creek: the lake water was comparatively cool in autumn and stable at about 4°C all winter, whereas the creek temperature was near freezing all winter. Oxygen levels were also a point of concern. Previous studies with incubating eggs in water from Auke Lake had shown that oxygen levels might drop to 8 mg/liter or lower (Bailey and Evans, 1971). EXPERIMENTAL DESIGN The plan of the experiment was to incubate pink salmon eggs at two different densities in four boxes of gravel and to compare the emergent fry with fry migrating seaward from natural redds in Auke Creek and with fry produced on flat trays of a Heath in- cubator. The evaluation of the incubator test was based on survival from eyed egg to emergent fry, fork length of preserved fry, wet weight of preserved fry, timing of emergence, and stage of development at emergence. WATER QUALITY Five parameters of water quality were monitored during the test of gravel incubation; viz temperature, oxygen, ammonia, carbon dioxide, and pH. Water temperatures of Auke Creek surface flow and the incubator effluent were read once daily to the nearest 0.1° C with a mercury thermometer. Dissolved oxygen values of samples collected at weekly intervals from the water supply and from the incubator effluents were measured to the nearest 0.01 mg/liter by the Winkler method. Water samples were collected at least once weekly from the incoming water supply and from the in- cubator effluents to analyze for ammonia, carbon dioxide, and pH. The samples were analyzed in con- nection with a laboratory bioassay (to be reported separately) of ammonia toxicity to salmonid eggs and alevins. "- ■■' *<^P^M other equipment needed to rear salmon fry. Four incubators are FigUre 2-Incubation^Ui^ ^sin^ravel JjjJ^"^ ^ seeded with pink saimon eggs. COLLECTION AND PRETREATMENT OF EGGS Pink salmon eggs from two sources, Auke Creek and Sashin Creek, were incubated in the eyed stage in flat trays and baskets so that eggs taken on different days could be mixed before they were seeded into test in- cubators. Eggs from Sashin Creek were used to fulfill the numbers needed for the experimental design. To prevent genetic contamination, no fry from Sashin Creek eggs were released at Auke Creek. Auke Creek Eggs Eggs for three of the four boxes were obtained from Auke Creek pink salmon. Adult pink salmon began entering the stream 12 August, and the run continued until 5 October. Eggs were collected from 159 females between 19 August and 15 September. We assumed an average egg content of 1,700 eggs per female. Females generally had to be held in a pen for about 5 days to allow their eggs to ripen enough so that they could be spawned artificially. The eggs were taken by incision after the females were killed by a blow on the head and bled by cutting the caudal artery; care was taken to avoid contamination of the eggs with blood, slime, or water. Three to five females were spawned into a plastic pail, and sperm from an equal number of males was added. The eggs and sperm were then gent- ly mixed and washed with fresh water. The flat trays and baskets in which the eggs were incubated to the eyed stage were covered to exclude light. Malachite green treatments (15 ppm for 1 h at 10-day intervals) were used to control fungus growth. All live eggs were eyed by 28 October at which time dead eggs were removed. The live eggs were seeded into the boxes of gravel 10 November (Fig. 2). Sample counts of eggs at the time they were buried in the gravel boxes indicated less than 1% were dead or not fertilized (opaque white or contained no embryo). Drip-drained Auke Creek eggs had an average count of 606 eggs/100 ml of displaced water. The displace- ment method of counting eggs is accurate to about ±5% (Burrows, 1951). Sashin Creek Eggs Eggs for the fourth box and the Heath incubator were obtained from Sashin Creek. The Sashin Creek eggs were taken 13-19 September by the same techniques described for the Auke Creek eggs, and were incubated to the eyed stage at Little Port Walter. They were transported to Auke Bay the first week of November. Dead eggs, 10.1% of the original total, were removed after eyeing. Sample counts of eyed eggs at the time they were buried in the gravel boxes indicated only 1.2% were dead or not fertilized. Drip-drained Sashin Creek eggs had an average count of 619 eggs/100 ml of displaced water. INCUBATOR DESIGN AND OPERATION The boxes were 1.2 by 0.91 by 0.91 m (4 by 3 by 3 feet) deep. The surface area of gravel was 1.1 m2, but each box occupied almost 1.5 m2 of floor space. The incubators containing gravel were of two types— Bams box and NMFS box — which differed only in the method of achieving uniform upwelling flow of water through the eggs and gravel. The method of dispersing water in the NMFS box required less space, and therefore more space was available for incubation of eggs and alevins in NMFS boxes than in Bams boxes. The two types of gravel incubators and the Heath incubator are designated in this report as Bams boxes A and C, NMFS boxes B and D, and tray incubator. Bams Boxes Water was introduced into the Bams box through a grid of pipes laid on the bottom inside the box. The grid of pipes in the Bams box (Figs. 3, 4) comprised a manifold pipe and eight pairs of cross pipes. The tJ MANIFOLD PIPE 1 .9 CM PVC CROSS PIPES Figure 3. — Bams box used for incubating pink salmon eggs in a gravel substrate. Grid of perforated pipes on bottom of box provides uniform distribution of upwelling flow. manifold pipe was 5.1-cm (2-inch) PVC, and the cross pipes were 42.5 cm long, 1.9-cm (3/4-inch) PVC placed on 15-cm centers. The cross pipes had 80 pairs of 2- mm (i,- inch) holes drilled 90° apart and were plac- ed holes up on 8-cm (3-inch) centers. The coarse gravel substrate in which the grid of pipes was buried and which was used as support for the eggs was river h till •WATER LEVEL 11 CM 5 CM COARSE SUBSTRATE ONE-SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE " 6TH LAYER ONE-SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE 2D LAYER ONE-SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE 1ST LAYER 8 CM BIRDSEYE GRAVEL 3.2 MM-6.4 MM DIAMETER PARTICLES 15 CM COARSE SUBSTRATE 1.3 CM-3.2 CM DIAMETER PARTICLES 2-MM PAIRED HOLES ON 8-CM CENTERS AND 45° FROM VERTICAL 1.90 CM PVC PIPE ■6.35 MM FIBERGLASS TANK WALL THICKNESS Figure 4. — Cross-section detail of Bams box used for incubating pink salmon eggs in gravel substrate (see Fig. 3). gravel that had been washed and graded to remove all particles under 1.3 cm (V2 inch) and over 3.2 cm (2V4 inch) in diameter. Crushed rock of 1.9- to 3.8-cm (3/i- to 1 ¥2 -inch) particle size is recommended but was not available. An 8- cm layer of bird's-eye gravel, 3.2- to 6.4-mm (Vs- to '/4-inch) particle diameter, acted as a pressure plate to achieve uniform water velocities in all parts of the box. The layer of bird's-eye gravel was also river gravel washed and graded to remove all par- ticles under 3.2 mm and over 6.4 mm in diameter. The boxes were loaded by placing the pipe grid on the bot- tom and packing a 15-cm layer of coarse substrate around and on top of it; this was followed by the 8-cm layer of bird's-eye gravel. Six 8-cm layers of coarse substrate were then placed in the box, and one-sixth of the total number of eggs was placed on top of each layer. The last layer of eggs was covered with 5 cm of coarse substrate. About 11 cm of open water was left at the top so that emergent fry could swim to the out- let on the side of the box. With this method of loading the box, about 0.57 m3 of coarse substrate habitat was accessible to the incubating alevins. Bams box A was loaded with 112,200 Sashin Creek eggs and box C with 53,600 Auke Creek eggs (Table 1). Water flow to the high-egg-density box (A) was 56 liters/min and to the low-density box (C) was 28 liters/min. Table 1. — Density of eggs, water flow, and apparent velocity of NMFS boxes B and D and Bams box C (Auke Creek pink sal- mon eggs) and Bams box A (Sashin Creek eggs). Bams boxes NMFS B boxes A C D Density: Eggs/box 112,200 53,600 56,100 112,100 Eggs/nr 100,600 50,300 50,300 100,600 Eggs/m3 198,000 94,700 74,200 148,400 Waterflow: Liters/min 56 28 28 56 Liters/min/10,000 eggs 5.0 5.2 5.0 5.0 Apparent velocity (cm'/h/cnr) 302 151 151 302 NMFS Boxes In the NMFS box, water was introduced through a false bottom of perforated plastic sheet stock (Figs. 5, 6). The false bottom was a sheet of perforated polypropylene plastic 3.2 mm {Vs inch) thick with 2.4- mm (| -inch) holes drilled on 12-mm ( | -inch) staggered centers. The perforated plate was supported on 19-mm (Vi-inch) diameter by 3-cm (1 r,U inch) long PVC rods on 7.6-cm centers. A 2.5-cm (1-inch) layer of bird's-eye gravel was placed on top of the per- forated plate to promote even distribution of water flow through the eggs and gravel. The bird's-eye gravel was covered with 15 cm (6 inches) of coarse substrate; then followed six 8-cm (3-inch) layers of coarse substrate. As with the Bams box, one-sixth of the total number of eggs was placed on top of each layer. The last layer of eggs and gravel was covered with 7 cm of coarse substrate. About 11 cm of open water was left at the top so that fry could swim to the outlet on the side of the box. Because the perforated plate system of flow dispersal occupied less space than the perforated grid system, the NMFS boxes had about 0.76 m3 of space available for incubating 5.1 CM PVC VALVE- 5.1 CM PVC UNION- 5.1 CM PVC PIPE- OUTLET, 5.1 CM PVC BULKHEAD FITTING ■PERFORATED POLYPROPYLENE PLATE (2.1 MM HOLES ON 12-MM CENTERS) Figure 5. — NMFS box used for incubating pink salmon eggs in a gravel substrate. Perforated false bottom provides a uniform distribution of upwelling water. alevins, compared with about 0.57 m3 in the Bams box. NMFS box B was loaded with 56,100 and box D with 112,100 Auke Creek eggs (Table 1). Water flow to the low-density box (B) was 28 liters/min and to the high-density box (D) was 56 liters/min. Tray Incubator The tray incubator (Heath) was seeded with 8,000 Sashin Creek eggs in each of two trays. The density was reduced to about 2,000 alevins per tray after hatching. Water flow was set at 19 liters/min. The fry produced on the flat screen trays without gravel were compared with fry from the gravel incubators and from the creek. The tray incubator also provided a source of fry to establish the basic parameters of an index to stage of development (Bams, 1970). NATURAL SPAWNING IN AUKE CREEK In 1971, 737 females and 1,035 males were placed above the weir and most spawned in the 297-m section of stream between the weir and Auke Lake, but a few spawned in Lake Creek above Auke Lake. The Auke Creek streambed above the weir has approximately 859 rrr suitable for spawning. A few pink salmon also spawned in the intertidal zone below the weir, but we did not estimate their number. ENUMERATION OF FRY Fry that emerged from the three boxes seeded with Auke Creek eggs (Table 1) were allowed to swim through parallel troughs that divided them into four groups. Three of the groups were allowed to escape immediately into Auke Creek, and the fourth group from each box was released the following night after enumeration and collection of samples of fry for size measurements. Total counts of fry that swam through each of the four partitioned passages on 12 successive C a tic ■WATER LEVEL 11 CM 7 CM COARSE SUBSTRATE */ ONE SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE 6TH LAYER ONE-SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE 2D LAYER ONE-SIXTH OF EGGS AND 8 CM OF COARSE SUBSTRATE 1ST LAYER ^ 15 CM COARSE SUBSTRATE a/ 1.3 CM-3.2 CM HAMETER PARTICLES 2.5 CM BIRDoEYE GRAVEL 3.2 MM-6.4 MM DIAMETER PARTICLES 3.2 MM PERFORATED POLYPROPYLENE PLATE 3-CM LONG SUPPORTS ON 7.6-CM CENTERS 6.35 MM FIBERGLASS TANK WALL THICKNESS Figure 6. — Cross-section detail of NMFS box used for incuba- ting pink salmon eggs in gravel substrate (see Fig. 5). nights were treated statistically as a two-way analysis of variance with one observation per cell. The calculated F value for passages (4.77) was significant at the 1% level with 3 and 33 df. Therefore, we con- cluded the four partitioned passages did not divide the migrating fry equally. We compensated for the ap- parent bias by retaining the fry from a different passage each night on a rotating basis. At the end of the migration, we calculated the total number of fry by two methods: first by assuming the nightly counts represented 25% of the migrating fry, and second by applying the estimates of sampler bias to the nightly counts. The two methods yielded estimates that agreed within 1%. Fry that emerged from the box seeded with Sashin Creek eggs (Table 1) were counted daily and destroyed after the samples were obtained for size measurements. The population of creek fry resulting from natural spawning was estimated by a hydraulic pump census (McNeil, 1964) of the streambed above and below the weir. We used a 0.1 m2 sampling frame. Sampling was physically difficult because of the presence of many boulders and outcroppings of bedrock. Standard error of mean was 45% of the value of the mean fry density. COLLECTING AND PROCESSING SAMPLES OF FRY To compare the size of fry from each type of incubation environment, samples of 50 fry were collected and preserved in 5% Formalin daily from each gravel incubator and as available from a migrant fry trap on Auke Creek. Samples of fry from the in- cubator tray were preserved weekly 21 March to 4 May 1972. Samples were kept in the preservative for 6 wk or more. The fry were then measured to the nearest millimeter of fork length and weighed to the nearest milligram of wet weight. The stage of development of emergent fry was described by Bams (1970) developmental index: W\ = proportion of run leaving in ith period. 10 V weight in milligrams D length in millimeters Because it was not practical to measure all of the preserved fry, representative samples from each in- cubator and from Auke Creek were selected about once each week. Weighted means and variances of pooled data were computed on the basis of the frac- tion of the migrant fry represented by each sample. Statistical comparisons were made of lengths, wet weights, and Bams (1970) developmental index as follows: Yw =ZWiYi, where YiV = weighted mean, Y{ = observed mean measurement in ith period, V(Y,„) = W' n V(Y where V{YW) = V(Ji) = estimate of weighted variance, sample variance of estimated mean in ith period, number of periods sampled. RESULTS The results of this study can be conveniently divided into those that measured and compared various parameters of the artificial and natural in- cubation environments and those that measured and compared emergence and morphometric features of the artificially and naturally produced pink salmon fry. Water Quality Water temperatures in the incubators and in the stream were recorded daily during the entire incuba- tion period. Dissolved oxygen, ammonia, carbon diox- ide, and pH of the influent (incoming water supply) and the effluents from the incubators were recorded at approximately weekly intervals from the beginning of hatching of the eggs until most of the fry had emerged. Temperature. — During the time the eggs were collected and fertilized (21 August to 15 September 1971), Auke Creek water temperature dropped from 17° to 12°C (63° to 54°F), and the incubator temperature dropped from 8.3° to 7.2° C (47° to 45°F) (Fig. 7). The incubator temperatures were lower than the creek because the water for the incubators came from below the surface of the lake and the creek water came from the surface of the lake. The incubator temperature remained near 7°C (45°F) until 18 Oc- tober and then gradually dropped to 3.5° to 3.8°C (38° to 39°F), where it remained from 16 November 1971 through 15 May 1972. Auke Creek temperature was higher in the fall and lower in the winter than in- cubator temperature. Auke Creek temperature dropped from 17°C at the beginning of the spawning season to the winter level of 0° to 1°C (32° to 34°F) in late November. Oxygen. — The oxygen level of the incubation effluents (Fig. 8) never dropped below 7.5 mg/liter (57% saturation) even though the incoming water supply was only 61 to 69% saturated with oxygen. Ox- ygen consumption rates were computed on the basis of numbers of fry emerging — about 0.02 mg of oxygen per fry per hour (Table 2). These estimates of oxygen comsumption rates are actually higher than the true rate for live alevins because dead eggs and alevins u °_12 - 1-10 < a. CREEK INCUBATORS ' -i 1 — t 1 { i~l — i — r — r 5 15 25 5 15 25 ' 5 1 — i — | 1 — i — i — i — i — i — I — i — i — i — i- i *~p''P' i 7" I I I I I I I I I I I I I I | I I I I I , I , . — i— 5 25 5 15 25 5 15 25 5 15 25 5 15 25 5 15 25 ' 5 15 25 5 15 25 AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER JANUARY FEBRUARY MARCH APRIL MAY Figure 7. — Temperatures of water in gravel incubators and of surface water in Auke Creek, 21 August 1971 to 15 May 1972. comsume oxygen (Brickell, 1971) and were not ac- counted for in the estimates. cr b'° rr a. a z -""N^ * ~~~~~~^^^ INFLUENT ■ - ° EFFLUENT {MEAN OF FOUR INCUBATORS) < J 5 0 1 ■ ■ ■ ■ .■ r— L '''■ | ' ' ' ' ' 1 ii.lillll.il 1 Figure 8. — Dissolved oxygen levels in the influent and effluents from four incubators at Auke Creek incubation facility, 1971-72. Influent value is from a single determination each date from main supply, and effluent value is the mean of four determina- tions— one from each of the four incubators. Table 2. — Fry production, flow rates, and oxygen consumption in four gravel incubators at Auke Creek during the period 1 February through 4 April 1972. Sediment. — A brown filamentous sediment ac- cumulated in all of the incubators during the incuba- tion period. The sediment was identified as precipitated iron compounds on and among the sheaths of iron bacteria. Fry Production and Quality The number of fry produced from Auke Creek eggs in our incubators was about equal to the production of fry from natural spawning in Auke Creek. We collected random samples of migrant fry from the in- cubators and from the creek to compare length, weight, and stage of development. Practically all of the fry migrated from the stream and from the in- cubators in April 1972. Except for fry needed to evaluate size and stage of development, fry from the Auke Creek eggs were released to enhance the natural run of pink salmon in Auke Creek. Survival from egg to fry. — We spawned 159 females and obtained 234,000 eggs or 86°~c of their es- timated contents of 270,000 (Table 3). Loss of eggs in this operation was attributed to attempts to take eggs Total Oxygen Number of Water oxygen consumed emergent flow consumed per fry Table 3. — Number of pink salmon eggs or fry at different life Incubator fry (liter/h) (mg/h) (mg/h) stages and survival between stages for Auke Creek eggs incu- Bams box bated in gravel incubators. A 109,100 48,300 3,360 1,680 1,633 842 0.0150 0.0174 Life stage Number present Percent survival C Eggs in females 270,300 — ~ NMFS box Green eggs 234,000 86 -69 B 50,500 1,680 1,039 0.0206 Eyed eggs 221,800 95 D 88,800 3.360 1,426 0.0161 Emergent fry 187,600 84 from females that were not ripe (i.e. eggs still at- tached to the ovarian membranes) and to shedding of eggs by females that had become overripe in the holding pen. Survival from newly fertilized eggs (green eggs) to eyed eggs was 95%. Losses during this interval depend on several factors: the ripeness of the fish; the care taken in handling the females, eggs, and sperm; and the growth of fungus among the in- cubating eggs. Typically, spawn takers expect 94 to 98% of the eggs they collect to survive to the eyed stage. Survival from the eyed egg to the emergent fry stage was 84%. A census of 36 random points with a hydraulic pump in April indicated that only 157,200 fry resulted from natural spawning in Auke Creek above the weir. The potential egg deposition (PED) of the 737 females was 1,252,900, so the survival was about 12.5% from PED to fry. No attempt was made to compare survival of fry from the tray incubator with survival of those from the gravel incubators. Effects of density on survival from egg to fry. — The density of eggs in the boxes ranged from 74,200 to 198,000 eggs/m3 of gravel and survival from eyed eggs to fry ranged from 79 to 97% (Table 4). Although the average survival in the gravel incubator was 89%, the inaccuracies inherent in the displace- ment method used to count eggs could amount to ±5% and only four incubators were tested. With this degree of accuracy and only two incubators in each test, only differences of more than 10% could be con- sidered significant. The cause of low survival in NMFS box D was not identified and could not be at- tributed entirely to the high density of eggs because Bams box A yielded good survival at the same densi- ty. Table 4. — Numbers of eyed pink salmon eggs and survival to emergent fry in individual incubators. Number of Numbr of Percentage Incubator eyed eggs emergent fry survival Bams box A 112,200 109,100 97 NMFS box B 56,100 50,500 90 Bams box C 53,600 48,300 90 NMFS box D 112,100 88,800 79 Effects of genetics and box design on survival from egg to fry. — Differences in survival due to genetic characters of parents or box design could not be conclusively demonstrated in this test because of the small number of samples. The one box of Sashin Creek eggs, Bams box A, had an eyed-egg-to-fry sur- vival of 90%, whereas three boxes of Auke Creek eggs had an average survival of 89%- (range 79 to 97%). Average survival from eyed egg to emergent fry was 90 to 97% in the two Bams boxes compared to 79 and 90% in the two NMFS boxes. The comparatively low sur- vival in NMFS boxes may have been only a chance oc- currence, and further tests should be conducted with this type of box. Effects of conditions in the incubators on size and stage of development of fry. — The completeness of absorption of the yolk by fry when they leave the in- cubating bed, either natural or artificial, bears direct- ly on the survival of the fry in the wild. Fry with a large amount of yolk have not attained their max- imum potential size, are relatively poor swimmers, and are unnecessarily vulnerable to predators. On the other hand, fry that have absorbed all of their yolk are losing weight and soon become weakened and emaciated and again unnecessarily vulnerable to predators. Fry emerge from gravel on their own voli- tion and presumably do so at the stage of develop- ment that ensures maximum survival. With tray in- cubators, the fry are locked in with screen covers until released. The hatchery operator must examine his tray fry and decide when to release them. In our experiment, the tray fry were not released, but based on measurements of samples preserved 21 March to 11 April 1972, they could logically have been released on about 21-28 March, when the tray fry had a mean Ku value of 1.97 (Table 5) which corresponded to the mean KD value of 1.97 of emergent fry from three of the four gravel incubators. Tray fry reached max- imum weight about 28 March. By 4 April, the tray fry were obviously becoming emaciated and losing vigor. Therefore, if maximum survival at sea had been the aim, they would have been released when they reached maximum weight, or a developmental index of about 1.97. Samples of tray fry collected 21 and 28 March had best pooled mean lengths, weights, and indices of development (Figs. 9, 10, 11); and we used these as a base to compare these factors in fry from the creek and gravel incubators. The source of eggs and incubation environment in- fluenced the size of fry, but the density of eggs did not. In gravel incubators, the Sashin Creek eggs yielded significantly longer fry than Auke Creek eggs (Fig. 9). With one exception, Sashin Creek eggs in gravel yield- ed significantly heavier fry (Fig. 10) than the Auke Creek eggs in gravel. Sashin Creek eggs incubated in gravel boxes yielded significantly longer and heavier fry than Sashin Creek eggs incubated in the trays. Creek fry were significantly longer than either gravel fry or tray fry regardless of egg source. Moreover, with the exception of Sashin Creek fry from Bams box A, creek fry were heavier than fry from incubators. Fry of Auke Creek parentage from the gravel incubators were significantly longer than fry of Sashin Creek parentage from the tray incubator, but in only one gravel incubator (Bams box C) were fry of Auke Creek Table 5.— Mean and variance of lengths, weights, and developmental indices of pink salmon fry from Auke Creek, gravel incubators ■• (Bams and NMFS boxes), and tray incubators. Length (mm) Weight (mg) Index Kjj Incubation site and date Number Mean Variance Mean Variance Mean Variance Auke Creek 11 April 50 32.00 0.9796 257.3 664.7 1.986 0.00166 18 April 19 32.55 1.6692 247.7 831.2 1.927 0.00265 22 April 49 32.96 0.8316 252.3 507.2 1.916 0.00169 26 April 49 32.65 0.9813 250.5 446.2 1.930 0.00190 5 May 50 32.08 1.7894 254.6 1,107.9 1.973 0.00300 17 May 50 32.44 0.6596 260.9 434.1 1.969 0.00272 Bams box A 30 March 50 31.56 1.3943 244.0 690.9 1.978 0.00266 5 April 50 32.06 0.8739 256.1 598.8 1.979 0.00198 11 April 50 32.18 1.1710 249.7 755.6 1.954 0.00156 18 April 50 32.08 0.5649 254.8 562.3 1.974 0.00190 22 April 18 31.98 0.9570 255.9 631.9 1.984 0.00188 26 April ;,o 31.82 0.9669 252.0 570.1 1.983 0.00130 5 May 50 32.04 0.8555 245.1 451.5 1.952 0.00284 Bams box C 30 March 50 31.40 1.3469 234.6 829.0 1.961 0.00186 5 April 50 31.32 1.2833 240.0 822.0 1.981 0.00214 11 April 50 32.00 0.9388 250.0 525.3 1.967 0.00157 18 April 50 31.86 1.2249 246.8 707.8 1.967 0.00167 22 April ;,() 31.84 1.0351 255.3 708.7 1.990 0.00196 26 April 1!) 32.06 0.6003 255.9 414.0 1.979 0.00164 5 May 50 31.86 0.6126 251.0 500.6 1.978 0.00263 NMFS box B 30 March 50 31.28 0.9812 235.2 643.6 1.971 0.00188 5 April 50 31.38 1.0159 236.6 512.4 1.969 0.00157 11 April 50 31.90 1.1531 247.2 514.6 1.966 0.00151 18 April 50 31.76 0.6759 243.4 508.2 1.964 0.00153 22 April 50 32.10 0.8265 254.2 367.5 1.973 0.00194 26 April 50 32.06 1.3228 252.8 601.9 1.971 0.00238 5 May 50 31.94 0.6698 245.0 415.0 1.958 0.00173 NMFS box D 30 March 50 31.44 1.1494 228.6 701.0 1.942 0.00149 5 April 50 31.60 1.0612 243.4 696.2 1.973 0.00170 11 April 50 31.80 1.0204 232.7 591.0 1.932 0.00154 18 April 50 32.20 0.8571 249.4 484.5 1.954 0.00164 22 April l:i 32.04 0.7483 254.7 397.0 1.977 0.00185 26 April 50 31.98 0.5914 243.9 381.7 1.952 0.00130 5 May 50 32.26 0.6045 250.5 373.7 1.953 0.00225 Tray 21 March 1!) 31.13 1.2801 233.7 824.0 1.976 0.00231 21 March 50 31.30 1.4388 234.1 772.8 1.966 0.00141 28 March 50 31.70 0.8309 246.0 626.3 1.974 0.00205 28 March 50 31.64 1.5004 241.4 996.0 1.964 0.00212 4 April 50 32.03 1.7698 234.8 1,165.9 1.921 0.00257 4 April 50 32.10 2.2551 235.8 1,242.3 1.921 0.00628 11 April 50 32.06 1.1175 227.7 941.3 1.901 0.00391 11 April 50 31.98 1.2853 227.7 941.3 1.906 0.00397 parentage heavier than fry of Sashin Creek parentage two high-density boxes, Bams box A and NMFS box from the tray incubator. The gravel incubator fry, D, were as long as or longer than fry from the two low- with the exception of fry from box D, had significantly density boxes, NMFS box B and Bams box C. The higher Ku indices (average 1.970 K{ 1 units) (Fig. 11) com paratively lighi :er weight of fry from high-density than creek fry (average 1.954 Kq units). Fry from the NMFS box D was probably relatable to the uniden- 10 SOURCE OF EGGS ENVIRONMENT AUKE CREEK N = 6 SASHIN CREEK N - 7 AUKE CREEK N = 7 AUKE CREEK N = 7 AUKE CREEK N = 7 SASHIN CREEK N = H 1 BAMS BOX A NMFS BOX B BAMS BOX C NMFS BOX D 31 30 31 40 31 50 31 60 31 70 31.80 31 90 32 00 32 10 32 20 32.30 32 10 32 50 LENGTH IMMI Figure 9. — Lengths of preserved fry in relation to source of eggs and incubation environment; the length of each horizontal line equals two times the standard deviation of pooled means from N samples of 50 fry per sample. SOURCE OF ECCS ENVIRONMENT AUKE CREEK N = 6 SASHIN CREEK N = 7 AUKE CREEK N = 7 AUKE CREEK N = 7 AUKE CREEK N = 7 BAMS BOX A NMFS BOX B BAMS BOX C NMFS BOX D SASHIN CREEK N = 1 210 2tS WEIGHT (MO Figure 10. — Weights of preserved fry in relation to source of eggs and incubation environment; the length of each horizontal line equals two times the standard deviation of pooled means from N samples of 50 fry per sample. SOURCE OF EGGS ENVIRONMENT AUKE CREEK N - 6 SASHIN CREEK N = 7 AUKE CREEK N = 7 AUKE CREEK N = 7 AUKE CREEK N = 7 CREEK BAMS BOX A NMFS BOX B BAMS BOX C NMFS BOX D SASHIN CREEK N = i) 1.915 1.950 1.960 1.965 Kr. INDEX 1.975 1.980 Figure 11.— Indices of development, KD, of preserved fry in re- lation to source of eggs and incubation environment; the length of each horizontal line equals two times the standard deviation of pooled means from N samples of 50 fry per sample. tified factor or factors that caused low survival in this box. The Kb index of alevins on trays decreased at a rate of 0.005 K/j units per day during the last month before hypothetical release of fry. If the Ku factor of alevins in gravel decreased at the same rate, then the gravel incubator fry migrated to sea at a developmen- tal stage 3 days younger than the creek fry. The alevins on trays were increasing in length at the rate of 0.08 mm/day during the last month before hypothetical release. If alevins in gravel were growing at a similar rate, 3 days additional growth would have increased their mean length from the observed 31.77 mm to 32.01 mm, which is still shorter than the 32.42- mm observed mean length of creek fry. Bams (1970 and 1972) produced gravel incubator fry that emerged prematurely and were slightly shorter than creek fry, but if the above correction was made for stage of development, Bams' incubator fry had the potential to equal creek fry in length. Time of emergence of incubator-reared and wild fry. — Fry emerged from the gravel incubators and migrated to Auke Bay between 23 March and 14 May; over 98f'r. of them left the incubators during the 34-day period between 30 March and 3 May. The midpoint of emergence occurred 10 April for NMFS box D, 14 April for NMFS box B, and 16 April for Bams boxes A and C (Fig. 12). These dates of emergence contrast sharply with the 21-28 March hypothetical best release dates for tray fry. The timing of migration of creek fry from Auke Creek could not be accurately determined because deep snow and ice delayed installation of a fry- counting trap. High streamflows resulting from rain and melting snow made it impossible to operate the trap continuously. The greatest trap count (680 fry) was recorded 11 April (Table 6), the first morning after the trap was in operation. By that time, 40% of the incubator fry had left the gravel boxes. A few creek fry were still leaving Auke Creek on 14 June, but heavy precipitation and high streamflows prevented further trapping. The trap catches suggest that tim- ing of creek fry migrations was generally similar to timing of incubator fry migrations but that creek fry continued to migrate over a much longer span of time. DISCUSSION This first attempt at mass production of high- quality pink salmon fry from gravel incubators at Auke Creek yielded 187,600 fry. The gravel incubators were evaluated on the basis of egg-to-fry survivals, size and stage of development in relation to conditions in the incubators, and time of emergence and seaward migration in relation to conditions in the estuary. i! y 55 f: ij I BAMS BOX A /,' I NMFS BOX D ll I BAMS BOX C NMFS BOX B T^^- T= I I I I I — i — i — i — I — i— i — i — I — i — i — I — I — I — i — I — I — r~ — r 23 27 31 2 6 10 11 18 22 26 30 6 10 11 MARCH | APRIL I MAY Figure 12. — Daily cumulative percentage of emergence of pink salmon fry from each gravel incubator at Auke Creek incuba- tion facility between 23 March and 14 May 1972. Survival Survival from PED to emergent fry was 69% for artificial incubation compared to 12.5% for natural spawning in the creek; a gain ratio (gain ratio = sur- vival in hatchery -f survival in creek) of 5.5. Gain ratios of 1.8 to 6.04 have been reported by Bams (1970, 1972, 1973) for pink salmon incubated in boxes of gravel in British Columbia. Approximately 181,000 incubator fry were released alive to migrate seaward with the 157,200 wild fry from Auke Creek. The remaining incubator fry, about 6,600, were used for size measurements. Eyed-egg-to-fry survival in the gravel incubators at Auke Creek was 84% compared to 82 to 95% survivals reported by Bams (1970, p. 1438; 1973, p. 4). This was encouraging because the water supply was not filtered at Auke Creek as it was in the British Columbia tests. The use of gravel incubators will be more practical if filtration of the water is not required. Size and Stage of Development Fry from the gravel incubators were intermediate in size between the larger creek fry and the smaller tray fry. Pink salmon fry emerged from the gravel in- cubators at an earlier stage of development, about 3 days, than from the Auke Creek streambed. We postulate that salmon fry actually find it easier to emerge from deep-gravel incubators which contain no sand than from streambeds which do contain sand. An additional 3-day development would not have Table 6. — Index trap fry migrating from catches of creek-incubated pink salmon Auke Creek, 11 April to 14 June 1972. Hours Number Hours Number Date fished caught Date fished caught April May 11 2000-2400 680 14 — — 12 2000-2400 250 15 — — 13 — — Hi — — 14 2000-2400 266 17 1630-0800 312 15 — — is 1630-0800 322 16 — — 19 1630-0800 532 17 — — 20 — — 18' 2030-2400 173 21 — — 19 2000-2400 321 22 — — 20 2000-2400 404 23 — — 21 2000-2400 205 24 — — 22 2000-2400 148 25 1630-0800 73 23 — — 26 1630-0800 26 24 2000-2400 203 27 — — 25 2000-2400 335 28 — — 26 2000-2400 244 29 — — 27 2000-2400 67 30 — — 28 2000-2400 130 31 — — 29 2000-2400 91 30 — — June May 1 — — 1 2100-0830 47 2 — — 2 2100-0800 58 3 — — 3 2100-0800 81 4 — — 4 2100-0800 122 5 — — 5 2100-0800 226 6 — — 6 7 2100-0800 147 7 8 — — 8 9 — — 5i 10 — — 10 — — 11 — — 11 — — 12 — — 12 — — 13 1630-0800 89 13 — — 14 1630-0800 33 12 made up for the difference in size between incubator fry and creek fry. The comparatively small size of fry from gravel incubators at Auke Creek must be at- tributable to our failure to provide a completely op- timum environment. At least three factors warrant further investigation and improvement at the Auke Creek Station: (1) substrate particle size and shape, (2) accumulation of brown sediment in the in- cubators, and (3) dissolved oxygen. Crushed rock was used by Bams (1970 and 1972) as the substrate for incubation of salmon eggs, whereas we used river gravel. The relative paucity of flat sur- faces of river gravel may have minimized the number of suitable interstices where alevins could find the physical support and undisturbed rest necessary for maximum growth. The brown sediment produced by iron bacteria in the incubators may have interfered with the uniform flow of water through the gravel interstices and physically restricted respiratory movements of gill opercula. Further study is needed to determine whether the brown sediment has a growth-inhibiting effect on salmon embryos. Dissolved oxygen in the effluents from the in- cubators exceeded 7.5 mg/liter, which is generally considered adequate for salmon. Percent of saturation values for oxygen in the influent ranged from 61 to 69% while percent of saturation values in the in- cubator effluents ranged from 57 to 68%. Higher ox- ygen levels might have resulted in larger fry. Itazawa (1971) found that rainbow trout 30 to 35 cm long in water at 2° to 9°C could not maintain normal arterial oxygen levels when oxygen concentration in the sur- rounding water dropped below 63% of saturation. He called this the "critical level of saturation" for rain- bow trout. Less oxygen would presumably constitute a stress which could suppress growth. Salmon alevins might have a critical level similar to rainbow trout. If this is true, then growth of alevins may have been in- hibited by low oxygen concentrations in this test of gravel incubation. Time of Emergence The time of emergence of fry from the gravel may be crucial to their survival. Survival in the estuary may depend on the arrival of fry coinciding with seasonal phytoplankton blooms, which could offer a measure of concealment from predators, or with seasonal zooplankton blooms, which are the major source of food. First spring phytoplankton blooms occur in Auke Bay each year between mid-March and early April, and zooplankton blooms soon follow (Wing and Reid, 1972). In 1972, zooplankton increased from scarce to relatively abundant sometime between 10 April 1972 and 18 May 1972 (B. L. Wing, Northwest Fisheries Center, Auke Bay Fish. Lab, NMFS, NOAA, Auke Bay, AK 99821, pers. commun.). Most of the incubator fry left Auke Creek to enter the es- tuary in April when zooplankton populations were in- creasing. Many of the creek fry migrated in April with the incubator fry, but some creek fry continued to migrate in May. ACKNOWLEDGMENTS The iron bacteria were identified by Joyce Gnagy, Biological Aid at the Auke Bay Fisheries Laboratory. LITERATURE CITED BAILEY, J. E., and D. R. EVANS. 1971. The low-temperature threshold for pink salmon eggs in relation to a proposed hydroelectric installation. Fish. Bull., U.S. 69:587-593. BAILEY, J. E., and W. R. HEARD. 1973. An improved incubator for salmonids and results of preliminary tests of its use. U.S. Dep. Commer., NOAA Tech. Memo. NMFS ABFL-1, 7 p. BAMS. R. A. 1970. Evaluation of a revised hatchery method tested on pink and chum salmon fry. J. Fish. Res. Board Can. 27: 1429-1452. 1972. A quantitative evaluation of survival to the adult stage and other characteristics of pink salmon (Oncor- hynchus gorbuscha) produced by a revised hatchery meth- od which simulates optimal natural conditions. J. Fish. Res. Board Can. 29:1151-1167. 1973. Evaluation of gravel incubators on first "hatchery" generation Tsolum River pink salmon, 1970-1972. Part I: Evaluation at the fry stage. Fish. Res. Board Can., Tech. Rep. 364, 18 p. BRICKELL, D. C. 1971. Oxygen consumption by dead pink salmon eggs in salmon spawning bed. M.S. Thesis, Univ. Alaska, College, 53 p. BURROWS, R. E. 1951. An evaluation of methods of egg enumeration. Prog. Fish-Cult. 13:79-85. ITAZAWA, Y. 1971. An estimation of the minimum level of dissolved oxy- gen in water required for normal life of fish. Bull. Jap. Soc. Sci. Fish. 37:273-276. MCNEIL, W. J. 1964. A method of measuring mortality of pink salmon eggs and larvae. U.S. Fish Wildl. Serv., Fish. Bull. 63:575-588. WING, B. L., and G. M. REID. 1972. Surface zooplankton from Auke Bay and vicinity, southeastern Alaska, August 1962 to January 1964. U.S. Dep. Commer., Natl. Oceanic Atmos. Admin., Natl. Mar. Fish. Serv., Data Rep. 72, 12 microfiche. ft U. S. GOVERNMENT PRINTING OFFICE: I974-697-; C8 /I4 REGION 10 13 PENN STATE UNIVERSITY LIBRARIES A0DDD72Dmn2