1. Groups of trout fry of the same parentage were grown in environments where the following factors were controlled : temperature, amount and intensity of illumination, rate of water flow, aeration and chemical composition of the water, amount of living space and quality of food supply. They were allowed to eat as much as they would, and individual weights were recorded during the first 8 months after the beginning of feeding.

  2. There was soon an increase in the range of individual weight in each group of fry, and thereafter the larger fry grew faster than smaller ones. When the larger fry were removed, the smaller ones grew at an increased specific rate, and when larger fry were added, the smaller ones grew more slowly. It is suggested that a ‘size hierarchy’ was established within each group, and an individual’s specific growth rate depended on its position in the order of decreasing weight.

  3. There was an optimum degree of crowding for maximum productivity. Compared with the fry in this group, the specific growth rates of individuals in larger, more crowded groups depended on the number of fish of larger size, while in smaller, less crowded groups, individuals grew at rates depending on the proportion of fish which were larger and smaller.

  4. Alevin weight had little effect on the specific growth rates of fry.

  5. There were differences between the growth histories of fry derived from alevins of the same weight and descended from the same father but different mothers (all of the same stock, age and size).

  6. The specific growth rates decreased as the fry grew older, but there was no correlation between body weight and specific growth rate, except for the size hierarchy effect within each group. This effect had a greater influence on the size of individual fry than had either alevin weight or heredity.

It has long been known that brown trout (Salmo trutta Linn.) grow to different sizes in different waters, and there has been much speculation concerning the factors determining these differences. Dahl (1918–19) suggested that the food supply and the degree of crowding were important; also that small, slowly growing trout were derived from smaller ova than those growing more rapidly. Southern (1932, 1935) demonstrated that, in Ireland, the rapidly growing fish are found in hard or alkaline waters, while those growing slowly live in soft, acid waters. This correlation has been upheld by the work of Frost (1939, 1945), Went & Frost (1942) and Went (1943) for Ireland, and by that of Raymond (1938) and Swynnerton & Worthington (1939) for Great Britain, though exceptions were reported in the Salmon and Freshwater Fisheries Report for 1936.

The greatest differences between the length and weight increments of trout living in hard and soft waters occur during the fourth year of life, but the annual specific growth rates show the most striking differences during the first year. After this, the average specific growth rates may be very similar in the two types of water although the fish are of markedly different sizes. They begin to spawn at an earlier age in some waters, and fish which spawn have a lower average specific growth rate for that year than fish of the same age which do not. Thus, the size attained by trout in different waters depends on (1) their rate of growth during the first year of life, (2) the age at which they begin to spawn, and (3) their average length of life. In soft waters, trout grow slowly during the first year, begin to spawn when 3 or 4 years old, and seldom live more than 5 years; in hard waters, they generally have a high specific growth rate during the first year, begin to spawn when older and may live for 12 years or longer.

The effect of environmental factors on trout growth is, therefore, of special interest during the first year of life, since there must be a factor or factors which inhibit growth in the soft waters during this year. A number of observations on trout growth are available from experimental hatcheries (McCay & Tunison, 1935, 1937) and from laboratories (Pentelow, 1939; Wingfield, 1940), but they concern fish more than 8 months old. It is generally admitted that a number of environmental factors may affect trout growth, but it has so far been impossible to correlate differences in growth rate with any particular factor, since several environmental factors have varied simultaneously without being measured accurately.

The present experiments were designed as an investigation of the rate of growth of individual trout in environments where as many factors as possible were controlled. Temperature, amount and intensity of illumination, chemical composition and aeration of the water, rate of water flow and quality of food supply were maintained as constant as possible, and the fish were allowed to eat as much as they would. Under these conditions it was hoped that the growth rates of the fish would vary in a definite and orderly way. Any changes in growth rate common to all the individuals should be the result of more or less intrinsic physiological factors, such as periodic secretion of endocrine organs. Any differences in growth rate exhibited by individuals in the same group might be ascribed to genetical differences or to the effect of the group on the individuals of which it is composed. When one environmental factor is varied and the others maintained constant, changes in the growth rate, compared with control fish, should indicate the effect of the varying factor.

The present paper is concerned with the growth of trout fry during the 8 months following the beginning of feeding. Although it is recognized that the conclusions derived from this study of the effect of genetical constitution, alevin weight, crowding and intraspecific relationships on specific growth rate are based on limited data, they are presented because they indicate some interesting possibilities.

Trout ova were obtained from The Midland Fishery, Nailsworth, Gloucestershire. On 10 December 1943, Mr Stevens stripped two 4-year-old brown trout, females A and B, of their ova, and fertilized each lot with milt from the same 4-year-old male fish. The fertilized ova were incubated at the hatchery in spring water at a temperature of about 10° C. The two sets of ‘eyed ova’ were sent to Cambridge by rail on 14 January 1944, and were there kept in troughs supplied with tap water at about 13° C. The majority of the ova obtained from female A had hatched by 21 January, while those from female B began to hatch on 20 January, and the majority had emerged by 24 January.

All the ova were dried on filter paper and weighed individually with a torsion balance before they hatched, and the alevins were then reared in hatchery trays in the dark. They were dried with filter paper and weighed individually when the yolk sac had nearly been absorbed and were then sorted into groups to be used for experiments. The fry descended from female A began to feed on 19 February, and those descended from female B on the 26th.

The weight distributions of the two populations of ova and alevins are shown in Fig. 1. In each case the alevins had a wider dispersion of weight and average weights which were 11 –12% higher than those of the ova from which they were derived (Table 1). The alevins descended from female A were used in the majority of experiments rather than those descended from female B, because the former were more numerous and their weight distribution curve approximated more closely to a normal curve.

Table 1.

Summary of data about ova and alevins

Summary of data about ova and alevins
Summary of data about ova and alevins
Fig. 1.

Histograms to show the distribution of weight among the populations of ova and alevins. A, descended from female A. B, descended from female B. Solid black: ova. White: alevins.

Fig. 1.

Histograms to show the distribution of weight among the populations of ova and alevins. A, descended from female A. B, descended from female B. Solid black: ova. White: alevins.

The alevins were grown in a constant-temperature room which was set at 11 ·5° C. This temperature was chosen because it lies within the optimum range for trout growth postulated by Pentelow (1939) and Wingfield (1940). The fish were kept in glass battery jars each containing 7 l. of Cambridge tap water, an analysis of which is given in Table 2. The water supply came from a constant-level tank so that it was under constant pressure. On entering the room the water flowed through a coil of metal tubing in the air and then through another coil immersed in standing water. This reduced its temperature to that of the room, and it seldom varied more than 0 ·2° from 11 ·5°C.

Table 2.

An analysis of Cambridge tap water (on 3 March 1944)

An analysis of Cambridge tap water (on 3 March 1944)
An analysis of Cambridge tap water (on 3 March 1944)

Each battery jar had its own water supply which sprayed through a jet and overflowed through a constant-level siphon into a gutter which ran to waste. The water in the jars was stirred continuously with compressed air. The room was free from daylight ; the fish were illuminated by 40 W. electric bulbs suspended 36 cm. above the water surface and so arranged that there was one bulb for each 20 sq.cm, of surface area. The lights were controlled by a clockwork time switch so that the fish were illuminated for 12 hr. and in darkness for 12 hr. each day.

The fry were fed on finely minced raw liver and were always given more than they would eat. The liver was dropped into the water, and the fry took it readily as it sank slowly to the bottom. At first they were fed twice daily, but after 3 months of feeding they were fed only once daily. The remains of food and faeces were siphoned out of the tanks everyday, but there was no attempt to measure the actual food consumption of the fry.

At intervals all the fry in each jar were removed; each fish was placed in 2 ·5 % urethane solution until it rolled over on to its side, and it was then dried on filter paper and weighed. Those weighing less than 350 mg. were weighed with a torsion balance to the nearest milligram, and heavier fish were weighed on a chemical balance to the nearest 5 mg. After weighing, the fry were placed in a shallow dish of water with ice floating in it, and they generally recovered within 2 min. There were very few deaths as a result of weighing; these were caused by the use of too strong a solution of urethane.

Some of the alevins descended from female A were sorted into five batches of fifty individuals, éach group with a limited range of individual weight, but covering between them the total range in weight of this family. Others were grown in similar environmental conditions but with different numbers of individuals in each tank. Two sets of fifty alevins descended from female B were also grown in the same environmental conditions. One of these groups included a range in weight which was the same as that of the majority of the alevins descended from female A, while the other consisted of individuals of greater weight.

At the beginning of the experiments some fry began to feed before the others and some never fed at all. The latter died within the first 3 weeks and others died at intervals later, but over 50 % of the individuals in each group survived for more than 8 months. The record for each tank showed a progressive decrease in the number of surviving fry, but only the smallest individuals died. Certain individuals were marked, and these remained in approximately the same numerical position if the weights of the fish were tabulated in descending order of size after each weighing.

It was decided not to use the mean weights for the groups in comparing the performance of fry in different tanks, since the weight distributions were not normal. The median value depended on the number of fry surviving at any date and would give no useful data for individual growth, though the medians and interquartile ranges indicated the dispersion of weight within a group. In order to estimate the growth of individuals, it has been assumed that each individual remained in the same position in order of descending weight for the group, so that, for instance, the values that were first, fifth, tenth, fifteenth and twentieth in successive weight records showed the change in weight of individuals occupying these positions in the order of size. With this assumption, individual specific growth rates were calculated from the formula
formula
where G = specific (or incremental, geometric, instantaneous or multiplicative) growth rate, YT = weight at time T, Yt = weight at time t, and T is later than t. The time was expressed as weeks, so that the specific growth rate is expressed as percentage weight per week.

When the growth records of the five groups of fifty fry descended from female A were compared, it was clear that the individuals did not all grow at the same specific rate, although they had the same parents and were living in the same environmental conditions. The dispersion of individual weights increased with time (Table 3), and individual specific growth rates were generally higher for the fish which grew most rapidly during the first period than for the smaller fry (Table 4). Thus, for the first few months, the order of descending weight was also an order of decreasing specific growth rate for the group. All the fish showed decrease in their specific growth rates as they grew older, and the difference in rate between the members of a group decreased with time.

Table 3.

Dispersion of individual weights within groups of trout fry

Dispersion of individual weights within groups of trout fry
Dispersion of individual weights within groups of trout fry
Table 4.

Mean specific growth rates (as percentage weight per week) of individual trout fry descended from female A

Mean specific growth rates (as percentage weight per week) of individual trout fry descended from female A
Mean specific growth rates (as percentage weight per week) of individual trout fry descended from female A

The effect of ‘grading’ on individual specific growth rates

The difference between the rates of growth of large and small individuals in the same group might be the result of two different causes: (1) some fry might be hereditarily incapable of as high a specific growth rate as others, since the individuals were not necessarily identical genetically although they were the offspring of the same parents and these came from an inbred hatchery stock; (2) the differences might result from some fish beginning to feed earlier than others and thus acquiring an initial advantage in size. Mutual reaction between fish of different sizes might then result in the larger fish growing faster and the smaller ones being retarded. Since the fish were allowed as much food as they would eat, competition for a limited food supply cannot be the factor involved.

In order to test this second hypothesis, two groups of fifty alevins descended from female A were grown under similar environmental conditions for 8 weeks, and all the fish in one group were marked by amputation of the adipose fin. The two groups were then combined, and the individuals were graded into equal numbers of ‘large’ and ‘small’ fry, i.e. those weighing more and less than 350 mg. respectively. These two groups were grown under the same environmental conditions as the original groups, and were weighed at intervals so that specific growth rates might be calculated.

Within each group, the larger individuals grew faster than the smaller fish, but the range in specific growth rate was approximately the same among the ‘large’ and ‘small’ fry (Fig. 2 and Table 5). The largest of the group of ‘small’ fry actually grew at a higher specific growth rate than the largest of the group of ‘large’ fry. The smallest fry of the ‘large’ group grew at a lower specific rate than the largest of the ‘small’ group, although before grading they had grown at nearly equal rates.

Table 5.

Comparison of the specific growth rates (as percentage weight per week) of large and small fry of the same age and with the same positions in the orders of size for their groups

Comparison of the specific growth rates (as percentage weight per week) of large and small fry of the same age and with the same positions in the orders of size for their groups
Comparison of the specific growth rates (as percentage weight per week) of large and small fry of the same age and with the same positions in the orders of size for their groups
Fig. 2.

The specific growth rates of four individual trout fry, all grown in the same environmental conditions. ○ This fish was the largest individual in its group throughout the experiment. Before week 8, the group consisted of large and small fry; after week 8, of large fry only. △ This fish was the twentieth in the order of decreasing weight for its group during the first 8 weeks. It then became the smallest individual in a group consisting of large fry only. ● This fish was the twenty-first in the order of decreasing weight for its group during the first 8 weeks. It then became the largest individual in a group consisting of small fry only. ▲ This fish was the smallest individual in its group throughout the experiment. Before week 8, the group consisted of large and small fry; after week 8, of small fry only.

Fig. 2.

The specific growth rates of four individual trout fry, all grown in the same environmental conditions. ○ This fish was the largest individual in its group throughout the experiment. Before week 8, the group consisted of large and small fry; after week 8, of large fry only. △ This fish was the twentieth in the order of decreasing weight for its group during the first 8 weeks. It then became the smallest individual in a group consisting of large fry only. ● This fish was the twenty-first in the order of decreasing weight for its group during the first 8 weeks. It then became the largest individual in a group consisting of small fry only. ▲ This fish was the smallest individual in its group throughout the experiment. Before week 8, the group consisted of large and small fry; after week 8, of small fry only.

After 7 months, the survivors of the two groups were transferred to another tank, and all the ‘large’ group were marked by amputation of the right, and all the ‘small’ group by amputation of the left pelvic fin. At this date, the range of weight in the two groups was nearly the same—from 770 to 8390 mg. in the ‘large’ and from 740 to 8150 mg. in the ‘small’ group. Of the fifteen largest fish, ten came from the ‘large’ and five from the ‘small’ group, while of the fifteen smallest fish, ten came from the ‘small’ and five from the ‘large’ group.

In the two months following this regrouping, the five largest fish of the ‘large’ set, now fish 1, 3,4, 5 and 6 of the mixed group, showed little change in average specific growth rate; the next five largest, now fish 7, 8, 11, 12 and 15, showed slight reduction in rate; but the five largest of the ‘small’ fish, now in positions 2, 9, 10, 13 and 14 of the mixed group, grew at a markedly lower average specific rate (Table 6). Fish with lower positions in the size orders also showed the effect of regrouping. Those in positions 7 –10 inclusive in the ‘small’ group and those in positions 16 –19 in the Targe’ group together filled positions 22 –29 inclusive in the mixed group. After regrouping, the ‘small’ fish grew at a reduced and the ‘large’ fish at an increased average specific rate so that the rates were nearly equal, although before regrouping the ‘small’ fish had been growing faster.

Table 6.

Comparison of the average specific growth rates of large and small fry for 10 weeks before and after regrouping

Comparison of the average specific growth rates of large and small fry for 10 weeks before and after regrouping
Comparison of the average specific growth rates of large and small fry for 10 weeks before and after regrouping

These observations all imply that the specific growth rate of an individual in a group of trout fry depends on its size relative to that of the other individuals, the larger fry growing faster than smaller ones. If the larger fry are removed, the smaller ones grow at an increased rate, and if larger fry are added, the smaller ones show a decrease in specific growth rate. There is no evidence that genetical factors play any important part in this difference in specific growth rate between individuals of the same family living in the same environment, though they may determine the readiness with which alevins begin to feed, and hence their initial specific growth rate.

The effect of crowding on individual specific growth rates

The growth rate of an individual in a group of trout fry might depend either on the total number of fish that are of larger size, or on the proportion of fish that are larger and smaller. Comparison of the growth of individuals in groups consisting of different numbers should make it possible to distinguish between these alternatives. Such groups, starting with 25, 50, 80, 100 and 150 alevins descended from female A, were grown in similar environmental conditions for 5 months after they began to feed. The mortality was consistently highest in the most crowded tank, but there was no marked difference in percentage survival between the four less crowded tanks. In every case, the fish which died were small and thin, but they did not appear to suffer from any definite disease.

In Fig. 3 the specific growth rates of the twenty largest fry in each group are compared. The values for the three larger, more crowded groups vary in the same way, those for the eighty alevin group being highest and those for the 150 alevin group lowest. If the values for the largest fish are excepted, the difference in growth rate among these fish is greater in the smaller, less crowded groups, especially in that started with twenty-five alevins.

Fig. 3.

Mean specific growth rates for 20 weeks after the beginning of feeding of the twenty largest individuals in groups containing different numbers of individuals.

Fig. 3.

Mean specific growth rates for 20 weeks after the beginning of feeding of the twenty largest individuals in groups containing different numbers of individuals.

Fig. 4 shows the specific growth rates of individuals occupying positions n/20, 2n/20, 3n/20, 4n/20, 5n/20, 6n/20, 7n/20, 8n/20, 9n/20 and 10n/20 in the size orders, where 71 is the original number of alevins in the group. The values for the three less crowded, smaller groups are very similar, while those for the two more crowded, larger groups cover a greater range of specific growth rate.

Fig. 4.

Mean specific growth rates for 20 weeks after the beginning of feeding of individuals occupying positions n/20, 2n/20, 3n/20, 4n/20, 5n/20, 6n/20, 7n/20, 8n/20, 9n/20 and 10n/20 (where n = number of alevins) in the size orders of groups containing different numbers of alevins.

Fig. 4.

Mean specific growth rates for 20 weeks after the beginning of feeding of individuals occupying positions n/20, 2n/20, 3n/20, 4n/20, 5n/20, 6n/20, 7n/20, 8n/20, 9n/20 and 10n/20 (where n = number of alevins) in the size orders of groups containing different numbers of alevins.

The fish in the tank started with eighty alevins grew at the highest specific rates in both cases, so this size of group and degree of crowding appear to be optimal for the environmental conditions used. Compared with this group, the specific growth rates of individuals in larger, more crowded groups depended on the number of fish of larger size, while individuals in smaller, less crowded groups grew at rates depending on the proportion of fish which were larger and smaller. Competition for a limited food supply can have had no effect, at least in the less crowded groups. The lower growth rates of smaller fry can be attributed to some ‘social organization’ within the group, those which began to feed earlier attaining a higher position in the hierarchy than the others and therefore continuing to grow at higher specific rates.

There was an optimum population density for which the yield of fish weight was maximum. In Fig. 5 the total weight of fish in each experimental tank is shown at intervals of 10 weeks. After 5 months the tank which was started with eighty alevins contained the greatest weight of fish, and the values decrease for degrees of crowding above and below this optimum.

Fig. 5.

Histograms showing the total weight of fish in groups of trout fry which were started with different numbers of alevins in equal volumes of water. Solid black: weight at the beginning of feeding. Cross-hatched: weight added during weeks 0 –10 after the beginning of feeding. White: weight added during weeks 10 –20 after the beginning of feeding.

Fig. 5.

Histograms showing the total weight of fish in groups of trout fry which were started with different numbers of alevins in equal volumes of water. Solid black: weight at the beginning of feeding. Cross-hatched: weight added during weeks 0 –10 after the beginning of feeding. White: weight added during weeks 10 –20 after the beginning of feeding.

The change in specific growth rate with time

The specific growth rates of most animals decrease with increasing age and size, and this occurs also with trout fry. The specific rate is highest during the first 3 weeks after the beginning of feeding, and the average rates decrease as the fish grow older(Table4), though there is some variation from week to week. Within each group the larger fish grew faster, but Table 5 shows that fish of the same age, but of very different sizes, grew at approximately the same rates if they occupied the same positions in the size orders of their groups. Thus, for trout fry, age is a significant factor determining the negative acceleration of specific growth rate with time, and there is no correlation between body weight and specific growth rate.

During the period between the hatching of the ova and the beginning of feeding, the alevins increased in weight by absorption of water (Gray, 1926), but the larger alevins were those which had hatched from larger ova. The three groups of fifty alevins descended from female A and the two groups of fifty alevins descended from female B, each group with a limited range of weight at the beginning of the experiments, thus represented a range in ovum weight.

During the first 8 months there was little difference between the average specific growth rates of fry derived from alevins of different weight, descended from female A, if individuals in the same positions in the size orders for their groups are compared (Table 7). Since, however, the alevins were of different initial weights, the larger ones grew into slightly larger fry than the smaller ones (Table 3).

Table 7.

Comparison of the growth of fry derived from alevins of different weights.

Comparison of the growth of fry derived from alevins of different weights.
Comparison of the growth of fry derived from alevins of different weights.

The two groups of alevins descended from female B did not include either the largest or the smallest alevins of their family, but differed from each other more markedly in weight than did any of those obtained from female A. The smaller alevins grew initially at a higher specific rate, so that the larger fry of the two groups were of approximately the same weight after 2 months. After this the fry derived from the larger alevins grew slightly faster, but at the end of 8 months, individuals in the same positions in the size orders did not differ markedly in weight (Table 7).

The groups of fry descended from female A and from female B thus showed different effects of alevin weight on early growth, but the range of specific growth rate within each group was greater than the differences between the rates of individuals of similar position in different groups. There were greater differences in size between larger and smaller fry in the same group than between fry of similar position, but derived from alevins of different weight.

All the fry used in these experiments had the same male parent, and their female parents, trout A and B, were of the same age, stock and size. The fifty smaller alevins descended from female B had a weight range which included that of one group of fifty alevins descended from female A. Since these two groups were grown in similar environmental conditions, any differences between their growth histories may be due to genetical factors inherited through the female parents.

The records for these two groups of fry show that their growth histories were not identical (Table 8). Those descended from female A had initially a higher specific growth rate than fish with the same position in the size order of the other group. During the third to eighth weeks, the two groups had approximately the same specific rates ; in the eighth to twentieth weeks, the fry descended from female B maintained these rates while those of the other group decreased; after the twentieth week, the rates of growth of the fry descended from female B fell to the same level as those of the others. Twenty weeks after the beginning of feeding, the fry descended from female B were markedly larger than individuals in the same positions in the size order of the group descended from female A, and they continued to be larger during the next 3 months until the end of the experiments.

Table 8.

Comparison of the specific growth rates (as percentage weight per week) of fry with the same male parent but different female parents

Comparison of the specific growth rates (as percentage weight per week) of fry with the same male parent but different female parents
Comparison of the specific growth rates (as percentage weight per week) of fry with the same male parent but different female parents

Within each group there was the same relationship between the size of an individual relative to the others and its specific growth rate, and the difference in size between large and small individuals in the same group was greater than that between individuals in the same positions in the size orders, but descended from different female parents. Thus, while the growth patterns of the two families were different, the factor which had the most marked influence on the specific growth rate of the individual fish was their position in the size order of their group.

Factors influencing the growth rates of animals may be of three types: genetical, physiological and environmental. Higgins (1929) and Davis (1934) showed that the average growth rate of Salvelinus fontinalis can be increased by selection of breeding stock for rapid growth, and they suggest that Salmo trutta and S. irideus are equally responsive to selection. The results of the present experiments showed that, even within the inbred stock of a hatchery, offspring of the same male but different female parents have somewhat different patterns of growth.

Minot (1890) pointed out that for most animals the specific growth rate is highest early in life and decreases, though with decreasing acceleration, as animals increase in age and size. This negative acceleration of specific growth rate depends on the age of trout fry, not their size, and may be partly a physiological effect of ageing of the tissues.

It is generally assumed that environmental factors are the most important of those concerned with the growth of fish. The principal physico-chemical factors are: temperature, amount and intensity of illumination, rate of water flow and concentration of gases and chemical substance in solution; the chief biotic factors are: quality and quantity of food and inter- and intraspecific relationships with other fish. Experimental work with salmonid fry has shown the importance of light (Tryon, 1942) and of rate of water flow (Washbourn, 1936). The effect of artificial foods on fry growth has been investigated in many hatcheries, and Coleman (1930) found that raw liver was the best food.

Gray (1928) and Wood (1932) showed the importance of temperature during the development of trout ova, and Gray (1926, 1928) also found a correlation between the embryonic specific growth rate of S. fario, the amount of yolk and the size of the embryo. Ova with less yolk grew into perfect fish of smaller size than normal. Dahl (1918 –19), Richmond (1924), Wilier, Quednau & Keller (1930) and Higgs (1942) all concluded that the size of the ova affects the growth rate of fry, larger ova giving rise to fry which grow faster. In the present experiments, alevin weight had little effect on specific growth rate. In one family, alevins of different sizes grew at the same average specific rates, so that the larger alevins became slightly larger fry than the others ; in the other family, the smaller alevins had initially a higher specific growth rate and grew into fry of the same size as those derived from the larger alevins. However, the range of aleyin weight used in these experiments was not so great as in those quoted above, and all the fish in the present experiments were closely related to each other. The average size of ova is probably characteristic of the female parent under natural conditions (Richmond, 1924) and may be associated with genetical factors which affect the growth rate of the fry.

The most important factor influencing the specific growth rates of individual fry proved to be the size of a fish relative to that of the others in the tank. Removal of larger fry resulted in increased specific growth rates of smaller ones, while addition of larger fry depressed the specific growth rates of smaller fry. This difference in growth rate, with the larger fish growing faster than smaller ones, was found in every experimental group. Since the fry were allowed to eat as much as they would, competition for food should not have been the factor producing this effect, and, indeed, the smaller fish appeared to feed as readily as larger ones. It seems possible that a ‘social hierarchy’, depending on size, soon developed within each group of trout fry, and that this influenced the specific growth rates of the individuals.

Observations on teleost fish with aggressive intraspecific behaviour have demonstrated that social hierarchies may exist in groups of fish, as they do in groups of other vertebrates such as hens (Schjelderup-Ebbe, 1922; Guhl & Allee, 1944). Noble and his collaborators (1938, 1939) have experimented with Xiphophorus, Hemichromis, Danio and other genera, while Braddock (1945) used Platypoecilus maculatus. When two individuals of these species meet they ‘challenge’ each other and one may bite or ‘nip’ the other. After several encounters one becomes ‘dominant’ to the other and will nip it without being nipped in return. Where several fish are kept together, they soon arrange themselves in a ‘nip order’, which may be a straight-line order or more complicated. Dominant fish tend to be larger than their subordinates, and may have better access to food and mates. They lose less weight during periods of starvation. Hierarchies of Platypoecilus last only 2 days on the average, but those of Xiphophorus may last for 2 or 3 weeks. Fish which normally swim in schools are less aggressive and their hierarchies are probably more complex and less easily investigated.

No aggressive behaviour comparable with nipping was observed among trout fry, but soon after they began to feed they developed the habit of remaining in definite positions in each tank unless disturbed. When some range in size had been established, it was possible to recognize individuals by their resting positions. Moon (1936) records similar behaviour among salmon parr in aquaria. Young trout fry are sometimes found in shoals in streams, and there is a range in specific growth rate among fry in the same stream. This can be shown by calculating their size at the end of the first year, using the annuli on the scales. For Straffan (River Liffey), for example, Went & Frost (1942) record a minimum length of 3 ·8 cm. and a maximum length of 14 ·2 cm. (the average being 8 ·6 cm.) at the end of the first year’s growth. This difference in growth rate might be the result of genetical factors, or of the selective action of a factor such as food supply, but it might also result from the existence of a size hierarchy in a stream, with correlation between the position of an individual and its specific growth rate. There is no evidence to indicate how the specific growth rate of a fish may be affected by its position in the hierarchy, either in streams or in tanks. In smaller, less crowded groups, however, an individual’s specific growth rate depends not only on the number of fish which are larger than itself, but on the proportion of fish which are larger and smaller.

There is also an effect of crowding on specific growth rates, and those fry which are more crowded than the optimum density have a lower level of growth rate. Retardation of growth by overcrowding may be caused by : (1) competition for a limited food supply, (2) mutual mechanical disturbance causing increased activity and thus increased food requirements, (3) accumulation in the water of excreta and other metabolic products. This last factor, the ‘conditioning of the medium’, occurs especially in confined spaces such as aquaria and fry ponds, and may be beneficial to some extent (Allee fit al. 1934, 1936, 1940). The first factor, competition for a limited food supply, should not have been operative in the present experiments, but the second or third factor might explain the retarded growth of the more crowded fish.

Pentelow (1944) suggested that the population density of trout is greater in soft than in hard waters, because of the superiority of spawning grounds and absence of predatory species of fish. Frost (1945), however, found no evidence in support of this suggestion for the River Liffey. The productivity of the limited volume of water in an aquarium is greatest for a medium population density, if the food supply is not limited, and increase in density leads to decrease in the level of specific growth rate. If the soft waters really are overcrowded with trout fry, this could be one factor contributing to the differences in specific growth rate between fry in the two types of water. Since, however, fry in hatcheries grow less rapidly in soft waters, there may be a direct effect of the chemical composition of the water on the growth of trout fry.

This work was carried out while holding a Research Studentship and later an Assistant Lectureship at Girton College, Cambridge. I should like to thank Prof. J. Gray, F.R.S., for his advice and encouragement throughout.

Allee
,
W. C.
,
Bowen
,
E. S.
,
Welty
,
J. C.
&
Obstino
,
R. B.
(
1934
).
J. Exp. Zool
.
68
,
183
.
Aller
,
W. C.
,
Finkel
,
A. J.
&
Hoskins
,
W. H.
(
1940
).
J. Exp. Zool
.
84
,
417
.
Allee
,
W. C.
,
Oesting
,
R. B.
&
Hoskins
,
W. H.
(
1936
).
Physiol. Zoöl
.
9
,
409
.
Braddock
,
J. C.
(
1945
).
Physiol. Zoöl
.
18
,
176
.
Coleman
,
G. E.
(
1930
).
Calif. Fish Game
,
16
,
1
.
Dahl
,
K.
(
1918-19
).
Salm. Trout Mag
.
17
,
18
. (English summary of Studies og Forsok over Örret og örretvand, 1917)
Davis
,
H. S.
(
1934
).
Trans. Amer. Fish. Soc
.
64
,
197
.
Frost
,
W. E.
(
1939
).
Proc. R. Irish Acad. B
,
45
, no.
7
.
Frost
,
W. E.
(
1945
).
Proc. R. Irish Acad. B
,
50
, no.
19
.
Gray
,
J.
(
1926
).
J. Exp. Biol
.
4
,
215
.
Gray
,
J.
(
1928
).
J. Exp. Biol
.
6
, 110,
125
.
Guhl
,
A. M.
&
Allee
,
W. C.
(
1944
).
Physiol Zoöl
.
17
,
320
.
Higgins
,
E.
(
1929
).
U.S. Dep. Commerce Fisheries Document
,
1096
.
Higgs
,
O.
(
1942
).
Salm. Trout Mag
.
106
,
216
.
Mccay
,
C. M.
&
Tunison
,
A. V.
(
1935
).
Rep. Cortland Hatchery
, no.
3
.
Mccay
,
C. M.
&
Tunison
,
A. V.
(
1937
).
Rep. Cortland Hatchery
, no.
5
.
Minot
,
C. S.
(
1890
).
Proc. Amer. Ass
.
39
,
21
.
Moon
,
H. P.
(
1936
).
Ann. Rep. Avon Biol. Res
.
1935
36
, P.
74
.
Noble
,
G. K.
(
1939
).
Amer. Nat
.
73
,
113
.
Noble
,
G. K.
&
Borne
,
R.
(
1938
).
Bull. Ecol. Soc. Amer
.
19
(
2
),
14
.
Noble
,
G. K.
&
Curtis
,
B.
(
1939
).
Bull. Amer. Mus. Nat. Hist
.
76
,
1
.
Pentelow
,
F. T. K.
(
1939
).
J. Exp. Biol
.
16
,
445
.
Pentelow
,
F. T. K.
(
1944
).
Nature, Lond
.,
153
,
464
.
Raymond
,
C. J.
(
1938
).
Salm. Trout Mag
.
93
,
329
.
Richmond
,
F. G.
(
1924
).
Salm. Trout Mag
.
31
,
48
.
Salmon and Freshwater Fisheries Report
(
1936
), p.
28
.
Schjblderup-Ebbe
,
T.
(
1922
).
Z. Psychol
.
88
,
225
.
Southern
,
R.
(
1932
).
Salm. Trout Mag
.
67
,
68
and 69.
Southern
,
R.
(
1935
).
Proc. R. Irish Acad
.
42
,
87
.
Swynnerton
,
G.
&
Worthington
,
E. B.
(
1939
).
Salm. Trout Mag
.
97
,
337
.
Tryon
,
C. A.
(
1942
).
Trans. Amer. Fish. Soc
.
72
,
145
.
Washbourn
,
R. M.
(
1936
).
J. Exp. Biol
.
13
,
145
.
Went
,
A. E. J.
(
1943
).
Proc. R. Irish Acad
.
49
,
269
.
Went
,
A. E. J.
&
Frost
,
W. E.
(
1942
).
Proc. R. Irish Acad
.
48
,
67
.
Willer
,
A.
,
Quednau
,
W.
&
Keller
,
A.
(
1930
).
Z. Fisch
.
28
,
565
.
Wingfield
,
C. A.
(
1940
).
J. Exp. Biol
.
17
,
435
.
Wood
,
A. H.
(
1932
).
J. Exp. Biol
.
9
,
271
.