Behavioral Ecology and Sociobiology

Behav Ecol Sociobiol(1982) 11:71-76

9 Springer-Verlag 1982

Interaction Between Schools of Fish and Mysids William N. McFarland i and Nancy M. Kotchian 2 1 Sectionof Ecologyand Systematics,Division of BiologicalSciences, CornellUniversity,Ithaca, New York 14853-0239, USA 2 Departmentof Biology,Universityof Maine, Farmington,Maine 04938, USA Received September 18, 1981 / Accepted June 11, 1982

Summary. A long-term daily census of 20 small patch reefs along the backreef of Tague Bay barrier reef, U.S. Virgin Islands, established that recently recruited postlarval French grunts, Haemulon flavolineatum, form mixed schools with at least two species of mysids (genus Mysidium). The formation of these taxonomically unrelated animals into socially interacting groups probably occurs because of the striking resemblance of postlarval grunts t o mysids. Over a period of 5 d and as the postlarval grunts grow, these mixed associations break down, only to reform every 14-15 d as new postlarval grunts periodically recruit from the plankton. The fleeting but persistent formation of these mixedspecies schools is believed to initially provide protection to the grunts and, as the grunts grow, limited protection to the mysids. In addition, the grunts benefit trophically from the associations because they prey on the mysids.

Materials and Methods Mixed associations of fish and mysids were observed along the backreef of the Tague Bay barrier reef at depths of 1 to 3 meters, and on a few occasionsin deeper locations (to 15 m) within the reef complex at Buck Island, St. Croix, U.S. Virgin Islands. All observations and collections of fish and mysids were obtained using SCUBA between March and October, 1980. Twenty separate small patch reefs (<10 m2), extending along 100 meters of the Tague Bay backreef, were flagged to census recruitmentof grunts (mostlythe French grunt, Haemulonflavolineatum) (McFarlandet al. in preparation). We repeatedly observed associations between postlarval grunts (PL's) and the schools of mysids(MYS's) on the census sites (Table 1). Five conditions were recognized and recorded on site visits (Table 2). Observationsincludeddescriptive data on the nature of the interactionbetween fish and mysids. The mysid species could not be identifiedin the field. Collections of mysid associations from the census sites, however, contained 3 homotypicschools of Mysidium integrum (Tattersall), 2 homotypic schools of M. columbiae (Zimmer), and 2 heterotypicschoolswith both speciespresent (Brattegard 1969). All identifiedpostlarval grunts were H. flavolineatum. Results

Introduction

The Social Unit - Daily Behaviors

Aggregations of animals, including schools of fish, flocks of birds, or herds of mammals, are often composed of more than one species. Typically, however, members of mixed-species groups are closely related (Wilson 1975) and this is especially so among vertebrates (for recent summary see Bertram 1978). A number of hypotheses account for mixed-species groupings, but two arguments are most persuasive-reduction of predation and improved foraging efficiency (fish-Eibl-Eibesfeldt 1962; Hobson 1968; Ehrlich and Ehrlich 1973; birds-Moynihan 1962; Morse 1967, 1970; Cody 1971). Here we report an instance where unrelated forms, mysid shrimps and fishes, school together.

Members of a mysid school are similar in size, utilize the reef structure for protection and, when threatened tighten ranks much like fish schools (Emery 1968). In Tague Bay mysid schools assume two general configurations: 1) compact schools of 10 to 1,000members which aggregate at topographic features on the reefs, such as coral overhangs and projections (Emery 1968, Fig. 2) or sea urchins (Diadema antillarum, Randall et al. 1964), and 2) extended schools of many thousands of individuals that spread diffusely over the sand bottoms between the small patch reefs. Both species of Mysidium occurred in each type of configuration.

0340-5443/82/0011/0071/$01.20

72

Fig. l a - d . Various views of the association of postlarval French grunts with mysid schools, a Overview of the association. At least nine PL's are visible, b Close up of PL's schooling with mysids. Note the overall superficial appearance between the two groups, e Enlargement of a PL and mysid schooling together, d Largest PL's schooling adjacent to a mysid school (see text and Table 1 for further details about the association)

Recently recruited PL's, which descend from the plankton and assume a semibenthic existence (McFarland 1980), often form mixed-species schools with MYS's. The PL's and MYS's in these associations look alike (Fig. 1) and are of simiIar size. PL's collected from mixed schools measured 8 to 13 mm TL and MYS's 8.3 to 10.4mm TL (tip rostrum to tail). Both PL's and MYS's feed actively on planktonic food throughout the day. Although PL's were most conspicuous when associated with compact MYS schools, we did observe some PL's within diffuse MYS schools over sand in the vicinity of the census sites; this associa-

tion may be more common than our impressions. Most of our census data therefore refer to the association of PL's with compact MYS schools.

Effect of PL Growth on the Association The social contact between PL's and MYS's is a fleeting one and spans a period of 5 or fewer days (Fig. 2). During 5 d a PL grows from 8.5 to 11.5 mm total length (Brothers and McFarland 1981), develops some dorsal body pigmentation (compare Fig. 1 a and c with d), and its schooling behavior begins to change (Table 1). The degree

73 I00

SITE 8

-

PL-2 (J3 >50 c.r

(M

3

5 7 9 JUNE 1980 i) SITE 19

EL d3 --

II

I0

Q_ o z

5

12

14

16 18 20 AUGUST 1980

2:>

Fig. 2. Turnover of postlarval French grunts in mysid schools on two census sites. Note that PL's are on a site for about 7 days. Arrows indicate the transformation of PL-I's to PL-2's. Half-moon symbols indicate dates of the third-quarter (upper panel) and first-quarter (lowerpanel) moons. See text for further detail

of social bonding between PL's and MYS's appears to depend most on their initial similarity in size and general appearance.

Evening Dispersion and Dawn Reaggregation of PL/MYS Schools Juvenile and adult grunts, which form protective schools on reefs during daytime, migrate and disperse to forage over grass or sand beds each night, and return to the same reef each morning (Ogden

and Ehrlich 1977; McFarland et al. 1979; McFarland and Hillis 1982). Because schools of postlarval grunts also disperse and regroup each day (Helfman et al. 1982) the PL/MYS schools must also reform each day. We summarize this break down and regrouping from 3 observation sets taken at census site 10, where 3 separate mixed-groups consistently formed. At approximately 15 min before sunset the MYS schools spread and moved slowly down the reef structure, the PL's still schooling with the mysids. From 8 to 10 min after sunset the various PL/MYS schools split as the mysids rose into the water column above the reef structure (Emery 1968; Hobson and Chess 1978). The PL's grouped tightly together and then quickly settled and spread over the bottom (ca. 10 to 12 min postsunset). The entire behavioral sequence observed at dusk is reversed at dawn. PL's appeared each morning over the sand where they were last observed the previous evening just a few minutes before the MYS's arrived. The PL's and MYS's formed a loose association at about 10 to 12 min before sunrise, and about 8 min before sunrise swam directly to the reef.

Population Characteristics of PL/MYS Schools PL's were never observed in association with MYS schools on two of the 20 census sites. We attribute this to the usual absence of compact MYS schools on these two sites. Although the census reveals that both MYS's and PL's were absent on 50% of all site visits, nevertheless, PL's were schooling with MYS's on 8% of all visits (Table 2).

Table 1. Developmental states of postlarval French grunts Development states

Size - SL mm (mean + 1SD/n/range)

Morphology

Behavior

PL-1

8.4+ 1.0/42/6.9- 9.9

Posterior body transparent with no distinct pigment marks

School as individuals with mysids, when threatened act like mysids. Eat very small mysids

PL-2

9.8 _ 0.7/21/8.5-11.2

Dorsal pigment marks caudal spot and vertebral skeleton distinct

School with other PL-2's within or on edge of mysid school, when threatened, conspecifically tighten ranks within mysid school. Prey on mysids

PL-3

12.3 _+2.4/48/8.9-18.7

Heavy dorsal pigment spots, body losing transparency, development of lateral melanistic stripe a

School with other PL-3's adjacent to mysid school, when threatened tighten their own ranks. Prey on mysids

The next life history state, the prejuvenile condition, is recognized by the development of horizontal lateral stripes and the development of migratory behavior (Helfman et al. 1982)

74 Table 2. Mean daily occupation of 20 schooling sites by various group combinations of postlarval grunts (PL) and mysids (MYS). The 20 sites were censused over 126 d between May 22 and October 24, 1980. Five group combinations were recorded: (1)PL and MYS absent; (2) PL present and MYS absent; (3) PL absent and MYS present; (4) both PL and MYS present, but not in mixed schools; and (5) both PL and MYS present in mixed schools Statistic

Type of group combination

(1)

(2)

(3)

(4)

(5)

PL(--) MYS(-)

PL(+) MYS(-)

PL(-) MYS(+)

PL(+) MYS(+) not mixed

PL(+) MYS(+) mixed

Mean daily occupation of all sites + 2 SE

63.3 + 1.1

22.9•

21.0•

8.3_+2.8

10.3_+4.0

% occupation

50.3

18.2

16.7

6.6

8.2

6O

PL's 2o4~ 0

0

I

5

,,,lllll,l,,,, 5

MYS's

I0

I0

15

15

50 0

I 2 34

5 6 7 8 9 101112131415

DAY OF PERIOD

Fig. 3. Frequency distribution of the occurrence of postlarval French grunt schools (PL's, upper graph), of PL/MYS schools (middle graph), and of mysid schools (lowergraph) for a 15 d semilunar period. Each histogram is the sum of occurrences on all census sites over 8 consecutive semilunar periods (see text for further details). The standard moon symbols indicate the times of full and new moons, and of the first and third quarter moons

Although the presence of P L / M Y S social units was highly variable from site to site, their formation was periodic. Larval French grunts recruit from the plankton with a semilunar rhythm with a period between 14 to 15 d (McFarland et al. in preparation). If dominated by recruitment of larval grunts, then the occurrence of the mixed P L / M Y S groupings should reflect a semilunar rhythm too.

To evaluate this possibility a frequency distribution for all the census data was generated by summing the daily records for all 20 sites from M a y 22 through September 18 (over 8 semilunar cycles). To approximate a semilunar cycle this summed time-series was wrapped repeatedly around a circle partitioned into 15 sections (see Fig. 3). Applying a Rayleigh test to these circular (--temporal) distributions (PL's, MYS's, and PL/ MYS) revealed that the PL's and the P L / M Y S associations were periodic, but the MYS schools were not (significant difference from a uniform distribution for Zo.99 = 4.6, for PL's Z = 26.4 and for P L / M Y S Z = 8 . 1 ; for the mysid schools Z = 0 . 7 5 , where Zo.95 =3.0, see Batschelet 1965). The semilunar periodicities of PL's alone and of PL/MYS's, however, were not in phase (Fig. 3). The number of PL/MYS groupings peaked about two days before the first or third quarter moon, while the peak for PL's coincided with the first and third quarter moons. This phase shift suggests that the earliest recruiting PL's were attracted to M Y S school sites or, alternatively, that later recruiting PL's found established PL schools more attractive than PL/ MYS or M Y S schooling sites. Stomach Contents of Postlarval Grunts Associating with Mysid Schools While photographing 6 small PL's schooling within a compact MYS school, and 9 large PL's (PL-1 and PL-3 in Table 1) schooling adjacent to the MYS's an unusual behavior was noted. Three to four of the larger PL's darted repeatedly into the mysid school. The fish were sufficiently small and the action so rapid that we could not detect what they were eating, although the MYS's seemed possible. The 15 postlarval grunts were chased into a d e a r plastic bag, preserved in 95% ethyl alcohol,

75

and their stomach contents examined. Two large PL's each contained a single large mysid and the smallest PL contained the cephalothorax of 3 very small mysids, clearly indicating that postlarval grunts can prey on mysids.

Discussion

The salient characteristic in the association between PL grunts and MYS's is it's transient nature. Their transparent appearance is likely the result of those selective pressures that adapt small, mobile, plankton-feeding animals for survival in planktonic communities. These mixed-schools are not limited exclusively to grunts and mysids for at greater depths they often include the masked goby, Coryphopterus personatus. What then might be the function of these mixed groups ? We suggest several functions. Initially the relationship is probably commensal, for the PL's may benefit most. At first formation (Fig. 2) there are few PL's in a group, and each PL is strikingly similar to it's MYS schoolmates. All the arguments that favor the anti-predator function of fish schools apply (Hobson 1979). But, do the mysids derive any benefit from the association? The answer is not obvious. Yet, once we knew what to look for we could quickly detect the presence of PL's in compact MYS schools-and we presume so could an experienced predator (see Milinski 1977a, b, 1979). Several reef fishes, and especially gobies in the genus Gobiosoma, feed on mysids (personal observations). Thus, MYS's may benefit if a predators attention is directed to the ever so slightly different PL's in their schools. Because the smallest PL's can feed on very small mysids the mixed-associations perhaps involves predation on newly hatched young from the ovigerous MYS's living within the mixed association (Steven 1961). Because the larger PL's prey upon their former mysid schoolmates, what perhaps begins as a commensal or even mutualistic relationship, certainly, also contains elements of a predator-prey system. The association of postlarval grunts with MYS's represents only one recruitment tactic in French grunts, for most PL's form large homotypic schools (McFarland et al. in preparation). Both tactics indicate that larval grunts descending from the plankton, or even before descent, develop a 'positive search-image' for objects similar to themselves (Fig. 1). It is suggested therefore that the identity of a grunts initial schoolmates lies mostly

in the domain of chance-for initial contact would be prescribed by first encounters and not by choice. Acknowledgements. We are indebted to Mr. Eldridge Birmingham for his dedicated service in the collection of data on postlarval grunts and their associations with mysids. Several divers, especially Richard Skorupski and Ms. Nancy Wolf provided indispensible help by acting as diving buddies. Dr. Thomas Bowman kindly identified several collections of mysids; Dr. George Dale first pointed out to us the mixed associations of gobies with mysids; Dr. Howard Howland assisted in the development of a BASIC program to analyze the data for periodicity; and Drs. Marjorie Reaka, Ted Hobson and Paul Sherman offered many valuable suggestions during preparation of the manuscript. The investigation was supported by NSF Research Grant OCE7918569. This paper is Contribution 69 from the West Indies Laboratory of Fairleigh Dickinson University.

References Batschelet E (1965) Statistical methods for the analysis of problems in animal orientation and certain biological rhythms. Monogr Am Inst Biol Sci, Washington, DC Bertram BCR (1978) Living in groups: Predators and prey. In: Krebs JB, Davies NB (eds) Behavioral ecology, an evohitionary approach, part 1. Blackwell, Oxford, pp 64-96 Brattegard T (1969) Marine biological investigations in the Bahamas. 10. Mysidacea from shallow water in the Bahamas and Southern Florida, part 1. Sarsia 39:17-106 Brothers EB, McFarland WN (1981) Correlations between otolith microstructure, growth, and life history transitions in newly recruited French grunts (Haemulon flavolineatum (Desmarest), Haemulidae). In: Lasker R, Sherman K (eds) Early life history offish, II. Raap R-v Reun Cons Int Explor Met 178 : 369-374 Cody ML (1971) Finch flocks in the Mohave Desert. Theor Popul Biol 2:142-148 Ehrlich PR, Ehrlich AH (1973) Co-evolution: heterotypic schooling in Caribbean reef fishes. Am Nat 107:157-160 Eibl-Eibesfeldt I (1962) Freiwasserbeobachtungen zur Deutung des Sehwarmverhaltens verschiedener Fische. Z Tierpsychol 19:165-182 Emery AR (1968) Preliminary observations on coral reef plankton. Limnol Oceangr 13:293-303 Helfman GS, Meyer JL, McFarland WN (1982) The ontogeny of twilight migration patterns in grunts (Pisces: Haemulidae). Anita Behav 30:317-326 Hobson ES (1968) Predatory behavior of some shore fishes in the Gulf of California. US Fish Wildl Serv, Res Rep 73 : 1-92 Hobson ES (1979) Interactions between piscivorous fishes and their prey. In: Clepper H (ed) Predator-prey systems in fisheries management. Sport Fish Inst, Washington, DC, pp 231-242 Hobson ES, Chess JR (1978) Trophic relationships among fishes and plankton in the lagoon of Enewetak Atoll, Marshall Islands. Fish Bull 76:133-153 McFarland WN (1980) Observations on recruitment in Haemulid fishes. Proc Gulf Carib Fish Inst 32rid annual session, pp 132-138 McFarland WN, Hillis ZM (1982) Observations on agonistic behavior between members of juvenile French and white

76 grunt schools-Family Haemulidae. Bull Mar Sci 32:255268 McFarland WN, Ogden JC, Lythgoe JN (1979) The influence of fight on the twilight migrations of grunts. Environ Biol Fish 4: 9-22 Milinski M (1977a) Experiments on the selection by predators against spatial oddity of their prey. Z Tierpsychol 43:311325 Milinski M (1977b) Do all members of a swarm suffer the same predation? Z Tierpsychol 45:373-388 Milinski M (1979) Can an experienced predator overcome the confusion of swarming prey more easily? Anim Behav 27:1122-1126 Morse DH (1967) Foraging relationships of brown-headed nuthatches and pine warblers. Ecology 48:94-103

Morse DH (1970) Ecological aspects of some mixed-species foraging flocks of birds. Ecol Monogr 40 : 119-168 Moynihan MH (1962) The organization and probable evolution of some mixed species flocks of Neotropical birds. Smithson Misc Collect 146 : 1-84 Ogden JC, Ehrlich PR (1977) The behavior of heterotypic resting schools of juvenile grunts (Pomadasyidae). Mar Biol 42: 273-280 Randall JE, Scfiroeder RE, Steak WA (1964) Notes on the biology of the echinoid Diadema antillarum. Caribb J Sci 4: 421-433 Steven DM (1961) Shoaling behavior in a mysid. Nature 192:280-281 Wilson EO (1975) Sociobiology. Belknap Press, Cambridge, Massachusetts