Aquaculture International 3, 60-65 (1995)

SHORT COMMUNICATION

Simple indices for the assessment of the

influences of social environment on growth
performance, exemplified by studies on Arctic
charr
Malcolm Jobling
NFH, University of Tromsq N-9037 Tromso, Norway

KEYWORDS:Arctic charr (.Sa/uelinus alpirzus), Dominance hierarchies, Feeding behaviour, Growth,

Social interactions

The majority of fish farmers and aquaculture researchers are aware that aggressive
behaviour and high levels of competition within a group of fish can have serious
consequences for growth. Nevertheless, only rarely have attempts been made to assess
the influences of social interactions upon the growth of farmed fish. The collection of
detailed information about social behaviour is time consuming, and the techniques used
(e.g. direct observation or video recording) are best suited for obtaining information
about individuals held in small groups, and so the study of the social environment of fish
has largely remained the domain of the behavioural ecologist (Thorpe and Huntingford,
1992; Grant, 1993).
If the fish within a rearing unit achieve rapid and homogeneous rates of growth, and
uniform body weights at harvest, it would be reasonable to conclude that the social
environment is favourable. Thus, high overall rates of gain in biomass, linked to uniform
rates of growth displayed by individuals and limited size disparity at harvest, might be
taken as being indicative of a ‘good social environment. Rapid and homogeneous growth
would, in turn, depend upon the food being evenly distributed among the fish in the
group.
Differences in the growth of fish within a group are often attributed to competition
among individuals for food; the more restricted or defensible the food supply, the greater
is the competition and the more pronounced is the disparity in food acquisition among
individuals (McCarthy et al., 1992; Grant, 1993; Jobling, 1995). Inequalities in food
acquisition result in disparity in growth and an increase in variation in size with time.
This has commonly been observed in small laboratory populations of fish provided with
restricted supplies (Davis and Olla, 1987; Thorpe and Huntingford, 1992; Grant, 1993).
The disparity in growth among individuals can be assessed by examining the coefficient
of variation (CV = [SD/ mean] x 100) of the size-frequency distribution; an increase in CV
with time indicates interindividual competition within the fish group. Thus, restricted
feeding is predicted to result in both low rates of biomass gain and increased size
disparity among the fish making up the group, and these predictions have been borne out
by the results of studies carried out on a range of salmonid species (Davis and Olla, 1987;
0967-6120 0 1995 Chapman & Hall
Social environment and growth 61
80
r l

I I I
0.5 1.0 1.5
SPECIFIC GROWTH RATE
FIG. 1. Relationship between group average specific growth rate and interindividual variability in
growth (CVsod for groups of Arctic charr fed at different levels of dietary restriction (0.5-l 5%
body weight day-‘). Specific growth rate = (in W, - In W,) x 100 / (f2 - t,), where W, and W’, are
fish weights at times t, and t2, and (t* - t,) is the time between weighings. Coefficient of variation,
CV = (SD / mean) x 100. Spearman rank correlation coefficient for the relationship between
group average specific growth rate and interindividual variability in growth rates: r = -0.929;
p < 0.01. From Jobling (1995).

McCarthy et al., 1992; Jobling, 1995). For example, Jobling (1995) reported that average
growth rates declined and growth variability (CVsc,) increased as amounts of feed
supplied to groups of Arctic charr, Salvelinus alpinus (L.), were reduced (Fig. 1.).
However, even in situations where the amount of food supplied does not appear to be
severely restricted, there may still be considerable variability in growth among
individuals making up a fish group (examples, Davis and Olla, 1987). Under such
circumstances an increase in size variation with time would indicate that behavioural
interactions can lead to growth suppression even in the absence of competition for
limited food supplies. For example, the large influence that behavioural interactions can
have on growth is demonstrated by consideration of the results obtained by Winberg et
al. (1992) in a study of the behaviour and growth of Arctic char-r held in small groups.
Dominant individuals, assessed on the basis of the numbers of aggressive acts initiated,
displayed positive rates of growth (Table 1: SGRsof fish ranked Al, Bl, Cl and Dl), whilst
many of the lower-ranking fish within each group lost weight (Table 1). Overall rates of
growth were very poor (mean SGRs -0.14-0.02% body weight day-‘) compared with
those that may be displayed by fish of similar size held together in larger groups (about
1% body weight day-’ for 100 g char-r at 10 “C) (Jobling et al., 1993). In addition, there was
a lack of precision in the data presented by Winberg et al. (1992). Precision is inversely
related to the spread among replicates, such that the greater the precision the greater
the reproducibility. In addition to poor overall growth and considerable intergroup
variability, there were found to be marked differences in rates of gain in weight among
individuals within each rearing group, with coefficients of variation ranging from 448 to
2164 (Table 1).
62 M. Jobling
TABLE 1. Influences of social status (dominance rank) on
the growth (SGR) of Arctic charr and the consequences of
the formation of dominance hierarchies on average rates of
growth and heterogeneity (CVsoa) within groups. SGR and
CV as defined in Fig. 1. Data from Winberg et al. (1992)

Fish rank SGRa Mean SGFP CVSGR

Al 0.62
A2 -0.18
A3 -0.43
A4 -0.39 -0.09 459
Bl 0.76
B2 -0.17
83 -0.04
B4 -0.47 0.02 2164
Cl 0.55
c2 0.33
c3 -1.05
c4 -0.40 -0.14 448
Dl 0.56
D2 -0.01
D3 -0.12
D4 -0.34 0.02 1589

a Units: per cent body weight day-‘.

Thus, hierarchy formation may lead to growth performance that would be considered
most undesirable under farming conditions - low rates of gain in biomass, inconsistency
of growth among groups, and marked heterogeneity in rates of gain in weight among
individuals within a group - and any manipulations that led to an improvement in the
social environment of the fish would be predicted to lead to increased growth. Changes
in social behaviour of Arctic char-r can be affected by manipulating stocking densities
and by exposing the fish to water currents, both forms of manipulation leading to a
decrease in the incidence of agonistic interactions (Christiansen and Jobling, 1990;
Brown et al., 1992).
Jorgensen et al. (1993) studied the effects of stocking density on feed intake and
growth in juvenile Arctic char-r. They reported that rates of both feed intake and growth
were higher in groups of fish held at high stocking density than in those reared at low
density (Table 2A: SGR) Examination of the data also suggests that the groups of fish
reared at the two higher stocking densities were more homogeneous than those held at
the lowest density (Table 2A: CVson),a highly significant negative correlation being found
between average growth rate (SGR) and growth variability (CV,,,) (Spearman rank
correlation coefficient, r = -0.983; p < 0.001). Not only were growth rates lower and
within-group variability greater, for the fish held at low density, but there were also
marked differences in group averages for fish growth among the different low stocking
density tanks (Table 2A). This inconsistency, or low level of precision, among tanks of
char-r held at low stocking density has also been observed in another recent study (Table
Social environment and growth 63
TABLE 2. Influences of stocking density and exposure to various water currents upon
average growth rates (SGR), growth heterogeneity (CVSGR)and interindividual variations
in feed intake (CV,) in groups of juvenile Arctic charr. Data from (A) Jorgensen et a/.
(1993), (B) Jorgensen and Jobling (1993), (C) Christiansen and Jobling (1990)

Stocking density
0% m”) n SGW CVSGR CVG

(4 15 18 0.45 151 136

15 20 0.07 704 212
15 19 -0.13 350 202
60 69 0.94 56 62
60 77 1.08 46 61
60 78 0.99 47 63
120 136 1.12 41 65
120 127 1.14 32 71
120 144 0.96 54 77

(8) 10 24 0.88 66 106

10 23 0.32 205 125
36 98 1.10 39 109
36 97 0.98 62 74
Current speed (cm s-‘) n SG!? CVSGR CVG

(Cl 0 20 0.631 98 63
0 19 0.175 394 120
6.5 22 0.848 56 45
6.5 23 0.371 174 80
13 12 0.863 63 59
13 24 0.825 67 76
19.5 24 0.994 63 69
19.5 8 0.721 63 57
26 22 1.623 39 65
26 23 1.263 45 42

a Units: per cent body weight day-'.

2B) (Jorgensen and Jobling, 1993). As stated previously, precision is considered to be

inversely related to the spread or dispersion among replicates, i.e. the greater the
precision the greater is the reproducibility or reliability.
In the study conducted by Christiansen and Jobling (1990), the levels of precision
increased in the groups of charr exposed to water currents flowing at higher speeds, and
increased precision was accompanied by both improvements in overall rates of growth
(Table 2C: SGR), and reduced heterogeneity of growth within groups (Table 2C: CVsc,).
There was found to be a highly significant negative correlation between group mean
SGR and CVscK (Spearman rank correlation coefficient, r = -0.861; p < 0.01). The
heterogeneous growth displayed by individual charr exposed to the lowest current
64 M. Jobling

speeds (Table 2C: CVso,) reflected inequalities in feeding, with large interindividual
differences in feed intake being recorded (Table 2C: CVo) (Spearman rank correlation
coefficient, r = 0.691; p < 0.05). Similar relationships between growth heterogeneity
(CV,,,) and inequalities in feed intake (CVo) have been recorded in stocking density
studies (Table 2A, Sperman rank correlation coefficient, r = 0.650; p < O.O5).These
results clearly suggest that social interactions may have considerable influences upon
the feed acquisition and growth performances displayed by individual fish.
Taking these findings as a starting point, it is possible to make some generalizations as
to how to assess social environment using simple observations of weight gain, size
variation and feed:gain ratios.
(1) A high rate of biomass gain (high average SGR)and little interindividual variability
in growth rate (i.e. low coefficient of variation for SGR),leading to little variation in body
size (i.e. low coefficient of variation for body weight), indicates a good rearing
environment (both biotic and abiotic) in which there is little or no depression of growth
due to agonistic behaviour or other forms of social interactions. Rapid and homogeneous
growth will generally be accompanied by the efficient utilization of feed for growth (i.e.
low feed:gain ratios).
(2) Poor and disparate growth (i.e. low average SGRand a high coefficient of variation
for either SGRor body weight) coupled to poor feed utilization (i.e. high feed:gain ratios)
reflects a poor social environment (e.g. possibly as a result of competition due to
underfeeding, or because of suboptimal temporal or spatial presentation of food).
(3) Poor growth with little variability indicates a general growth depression resulting
from a poor rearing environment (e.g. poor water quality), and a high feed:gain ratio
would reflect feed waste.
In the context of laboratory growth studies it is disconcerting to find that experimental
designs may be such that there is a considerable risk that social interactions among fish
within the rearing group may play a major role in influencing the outcome of the
experiment. For example, studies of fish nutrition have often been conducted using small
numbers of fish fed according to restricted feeding regimes under which they are
supplied with a fixed weight of feed, corresponding to a predetermined proportion of
their body weight, each day. These are rearing conditions under which there is likely to
be competition for the limited food resource, leading to disparity in food acquisition,
with the consequences of low overall rates of growth, and considerable heterogeneity
within groups. Because these are also conditions under which there may be low
precision, i.e. considerable spread or dispersion among replicates, it is clear that large
numbers of replicates will be required if reliable data are to be collected. Unfortunately,
few fish nutritionists are able to conduct experiments with more than 3-4 replicates per
treatment, which means that there is a risk that a type II error will be made when data are
analysed with respect to treatment effects, i.e. real treatment effects may be missed due
to the failure to reject a false null hypothesis.

REFERENCES
Brown, G.E.,Brown, JA. and Srivastava, R.K.(1992) The effect of stocking density on the behaviour
of Arctic charr (Salvelinus alpinus L.). Journal of Fish Biology 41, 955-963.
Christiansen, [Link] Jobling, M. (1990) The behaviour and the relationship between food intake
Social environment and growth 65
and growth of juvenile Arctic charr, Salvelinus alpinus L., subjected to sustained exercise.
Canadian Journal of Zoology 68, 2185-2191.
Davis, M.W. and Olla, B.L. (1987) Aggression and variation in growth of chum salmon
(Oncorhynchus keta) juveniles in seawater: effects of limited rations. Canadian Journal of
Fisheries and Aquatic Sciences 44, 192-197.
Grant, J.W.A. (1993) Whether or not to defend? The influence of resource distribution. Marine
Behaviour and Physiology 23, 137-153.
Jobling, M. (1995) Feeding of charr in relation to aquaculture. Nordic Journal of Freshwater
Research (in press)
Jobling, M., Jorgensen, E.H.,Arnesen, A.M. and Ring@,E. (1993) Feeding, growth and environmental
requirements of Arctic charr: a review of aquaculture potential. Aquaculture tnternational 1,
20-46.
Jorgensen, E.H. and Jobling, M. (1993) Feeding in darkness eliminates density-dependent growth
suppression in Arctic charr. Aquaculture International 1, 90-93.
Jorgensen, E.H., Christiansen, [Link] Jobling, M. (1993) Effects of stocking density on food intake,
growth performance and oxygen consumption in Arctic charr (Salvelinus alpinus).
Aquaculture 110, 191-204.
McCarthy, I.D., Carter, C.G. and Houlihan, D.F. (1992) The effect of feeding hierarchy on individual
variability in daily feeding of rainbow trout, Oncorhynchus mykiss (Walbaum). Journal ofFish
Biology 41, 257-263.
Thorpe, [Link] Huntingford, F.A. (eds) (1992) The Importance of Feeding Behavior for the Efficient
Culture of Salmonid Fishes. World Aquaculture Society: Baton Rouge.
Winberg, S., Nilsson, G.E. and Olsen, K.H. (1992) Changes in brain serotonergic activity during
hierarchic behavior in Arctic charr (Salvelinus afpinus L.) are socially induced. Journal of
Comparative Physiology 17OA, 93-99.