Factors affecting the reproductive success of common tern (Sterna hirundo) colonies on the lower Great Lakes during the summer of 1972

Transcription

1 Factors affecting the reproductive success of common tern (Sterna hirundo) colonies on the lower Great Lakes during the summer of 1972 RALPH D. MORRIS, RODGER A. HUNTER, AND JAMES F. MCELMAN Department of Biological Sciences, Brock UniversiQ, St. Catharines, Ont., Canada L2S 3A1 Received June 3, 1976 MORRIS, R. D., R. A. HUNTER, and J. F. MCELMAN Factors affecting the reproductive success of common tern (Sterna hirundo) colonies on the lower Great Lakes during the summer of Can. J. Zool. 54: The breeding biology of five common tern (Sterna hirundo) colonies in the lower Great Lakes was studied between May and August Frequent visits were made to each colony and data collected on nests within large artificially enclosed areas. The objective was to identify factors that influence the reproductive success of common tern colonies on the Great Lakes. Hatching success was significantly dependent on clutch size and time of clutch initiation, whereas fledging success was independent of clutch size. The most common category of egg failure was disappearance from the nest. One colony (Port Colborne) realized a significantly higher hatching and fledging success than the others, among which there were no significant differences in prehatch or posthatch success rates. Factors that contributed to differences in reproductive success are presented and their relative contributions to the reproductive success of the tern colonies are discussed. The factors include the relative proportion of three-egg clutches, incubation time as a measure of 'parent attentiveness,' numerical size of the colony, predation, competition for nesting sites by gulls, food availability, flooding, and toxic chemicals. We conclude that no single factor can be readily correlated with reproductive success and suggest that caution should be exercised when considering the relationships between reproductive success and factors influencing it. MORRIS, R. D., R. A. HUNTER et J. F. MCELMAN Factors affecting the reproductive success of common tern (Sterna hirundo) colonies on the lower Great Lakes during the summer of Can. J. Zool. 54: On a etudie la biologie de la reproduction chez cinq colonies de sternes (Sterna hirundo) dans la region basse des Grands Lacs, de mai a aoiit Les colonies ont 6tt visitkes souvent: on a recueilli des donntes sur la nidation, a l'interieur d'aires delimitees artificiellement. I1 s'agissait d'identifier les facteurs responsables du succes de la reproduction chez les colonies des Grands Lacs. Le succes de l'eclosion depend etroitement de la taille de la portke et du moment ou les ceufs sont pondus, alors que le succes de la phase qui suit l'eclosion jusqu'a la pousste des plumes ne depend pas de la taille de la nichee. La cause la plus frequente de I'insucces du stade ceuf est la disparition des ceufs du nid. Chez l'une des colonies (Port Colborne), on a remarqut que le succ~s de l'eclosion et celui de la phase qui va jusqu'au stade oisillon etaient significativement plus grands que chez les autres colonies ou on n'a pas enregistre de differences significatives dans les taux de succes avant l'eclosion ou apres l'eclosion. On presente ici les facteurs qui regissent le succes de la reproduction, de mzme que la contribution relative de chacun de ces facteurs au succts de la reproduction, chez les diverses colonies de sternes etudiees. Ces facteurs comprennent, entre autres, la proportion relative de nichees de trois oeufs, la duree de I'incubation consideree comme indice d' "attention" de la part des parents, le nombre d'oiseaux dans la colonie, la predation, la competition pour les sites de nidation apportte par la pksence de goklands, la disponibilitk de la nourriture, les inondations et la presence de produits chimiques toxiques. I1 ne semble pas qu'un facteur en particulier soit responsable du succes de la reproduction; il faut donc beaucoup de prudence lorsqu'on essaie d'etablir un lien entre le succes de la reproduction et les facteurs qui I'influencent. [Traduit par le journal] Introduction Since the summer of 1972 we have engaged in intensive studies to assess the significance of a variety of environmental factors on the breeding biology of common tern (Sterna hirundo) colonies on the lower Canadian Great Lakes. In North America long-term studies of common terns have been largely restricted to colonies along the eastern coast of the United States (e.g. Austin and Austin 1956; Nisbet and Drury 1972). No comparable data are available for common tern colonies in the Great Lakes. Common terns exhibit a remarkable range of variation in reproductive success (e.g. Howard 1968; Ludwig 1962; Nisbet and Drury 1972). Hatching success, usually expressed as the number of chicks hatched per egg laid, varied from almost total failure (Austin 1942; Marshall 1942)

2 MORRIS ET AL to almost complete success (Nisbet and Drury 1972). Fledging success is defined in the literature on a per egg laid, per egg hatched, or per nest basis, and also varies over a considerable range. The ultimate reproductive success of a tern colony may be influenced by a complex variety of factors. Several factors are known or suspected to reproduce the annual production of chicks by contributing to reproductive failure before the egg hatches (prehatch mortality) or to failure between the time of hatching and fledging (posthatch mortality). These include predation pressure (Paynter 1949; Hunter and Morris 1976), food availability (Lemmetyinen 1973; Nisbet 1973), competition for nesting sites by gulls (Morris and Hunter 1976a), flooding (Ludwig 1962; Nisbet 1972), incubation attentiveness (Morris and Hunter 1976b), and toxic chemicals (Gilbertson et al. 1976). In addition, the characteristics of the adult birds and the age structure of a colony can influence reproductive success in a particular year. For example, young or inexperienced common terns nesting for the first time often begin nesting late in the season and produce smaller clutches with a lowered hatching rate than their older, more experienced counterparts (Langham 1968; LeCroy and Collins 1972). It is the purpose of this paper (1) to report population demographic data collected from common tern colonies at three sites in the lower Great Lakes in 1972, (2) to assess the significance of some environmental factors that have a substantial influence on the reproductive success of common tern colonies, and (3) to suggest techniques and experiments that might be used to accurately measure the realized reproductive success of common tern colonies. The Study Areas Three common tern colony sites were selected, one in eastern Lake Erie near Port Colborne, Ontario, one in western Lake Ontario near Hamilton, Ontario (Hamilton.Harbor), and one on Muggs Island in Lake Ontario offshore from Toronto, Ontario. Site selection was based on variation in colony size, variation in a number of physical and biological factors at each colony, and the availability of some background information for the Hamilton Harbor site (M. Gilbertson, personal communication). The Port Colborne site was a chipped concrete shelf about 7 m wide and 300 m long, part of a man-made break wall associated with the Lake Erie terminus of the Welland Canal. Vegetation was sparse and widely scattered except in certain areas where a substantial grassy mat predominated. The total break wall area extended to the south and west of the tern colony and was intermittently used by fishermen and naturalists during the summer months. The Hamilton Harbor colony was on two small artificial islands formed by loosely piled, limestone rock, which partially supported electric power-line towers. The islands, designated Neare Island and Farr Island, were about 30 m x 35 m each. Neare Island was free of vegetation while Farr Island contained a single small tree surrounded by clumps of grasses, nettles, and thistles. Public access to these islands was completely restricted by the Ontario Hydro Commission. Muggs Island is one of a chain of partially artificial islands, most of which receive extensive commercial and recreational use during the summer months. The common tern colony was on the eastern end of Muggs Island on the edge of a sandy hillock. One section of the colony (Muggs I) was well covered by grasses and thistles while the other (Muggs 11) was on a sand substrate about 50 m away. As the site was designated a bird sanctuary, public access was controlled. Ground-nesting species at the common tern colonies included ring-billed gulls (L. delawarensis) and herring gulls (L. argentatus) at Port Colborne and Muggs Island. An estimated 1000 ring-billed gulls and 100 herring gulls nested at the Port Colborne site. Both species were spatially separated from the tern colony on elevated piles of loosely grouped limestone rock, above and to the west (Morris and Hunter 1976~). The ring-billed gull colony at Muggs Island, estimated at about individuals, occupied the major portion of the sandy hillock with one edge of the colony partially contiguous with the common tern nesting area. Eleven herring gull nests were widely scattered over the area but were always spatially separated from the ring-billed gull and common tern colonies. Methods All colonies were visited daily or every 2nd day from early May through late August Visits were at regular times each day to confine disturbance of each colony to a predictable interval. During nesting, incubation, and early feeding stages, our presence appeared to cause little

3 1852 CAN. J. ZOOL. VOL. 54, 1976 disturbance as adults would quickly return to the nest and young chicks seldom left the nest site. As the chicks matured some left the nest site and were occasionally attacked by neighbouring adult terns. At each visit records were kept of new nest starts, eggs laid and lost, hatching success, and subsequent fate of chicks. Nests were individually marked with wooden tongue depressors or by numbered paint marks on nearby rocks. Eggs were individually numbered with non-toxic pencils as they were laid and were handled on each visit. Thus, nonviable eggs or eggs that were no longer being incubated could be readily detected. Chicks were banded within 2 days of their hatching dates and searches on each visit established death or disappearaoce dates for each ringed chick within the enclosed area (see below). In early May 1972, substantial portions of each colony were fenced with chicken-wire meshing (2.5 cm hexagonal hole size). The wire meshing was buried in the substrate and supported at regular intervals by metal rods. The minimum height of the fences was nowhere less than 30 cm. This procedure provides an accurate assessment of fledging success as chicks cannot leave the enclosed area until they are able to fly (LeCroy and Collins 1972; Nisbet and Drury 1972). Details on the size of each enclosed area are in Table 1. There were two enclosed sites on Muggs Island, hereafter referred to as Muggs I and Muggs 11. The Muggs I1 colony was fenced on 24 June 1972 after the loss by water erosion of about 407, of the nests inside the Muggs I enclosure at about that time. The Hamilton Harbor islands were not fenced as both could readily be searched by an investigator and were already 'enclosed' by the surrounding water barrier. Small numbers of first eggs from three-egg clutches were collected from Muggs Island and Port Colborne from a part of each colony immediately adjacent to the fenced areas. Eggs were analyzed for organochlorine substances by Dr. L. Reynolds of the Ontario Research Foundation, according to the methods outlined in Reynolds (1969). Results Despite the geographic separation of the colonies and the variation in numbers of nesting pairs, there was a degree of synchrony in the dates of egg laying and egg hatching among the colonies (Table 2). The Muggs I1 colony was delayed because of the late start in egg laying, TABLE 1. The study areas and enclosure information Approximate size Nests of enclosed area, enclosed, Colony site Lake m2 n Port Colborne Erie Neare Island Ontario Farr Island Ontario Muggs I Ontario * Muggs I1 Ontario 'Detailed clutch size and hatching records kept on 111. although total egg production was 90% complete within a few days of the other colonies. Hatching Success There was a direct relationship between hatching success and clutch size with larger clutches being more successful than smaller clutches (Table 3). A 2 x 2 contingency table analysis with the Yates correction for continuity (x2,) revealed that at the Port Colborne, Muggs 11, and Farr Island colonies, three-egg clutches hatched significantly better than two-egg clutches (x2, = 55.3, 12.4, and 4.3 respectively). There were no differences at the Neare Island and Muggs I colonies (x2, = 1.36 and respectively). Langham (1968) and LeCroy and Collins (1972) reported similar observations on the influence of clutch size on hatching success in common terns. Hatching success was also related to the time of clutch initiation. Egg laying in the Port Colborne, Neare Island, and Farr Island colonies was 50x complete between 22 and 25 May 1972 (Table 2). Accordingly, data from the Neare and Farr Island colonies were divided into 'early' and 'late' nesters around a split date of 25 May 1972, whereas, data from the much larger Port Colborne colony were separated into three date groupings (Table 4). 'Early' eggs hatched significantly better than 'late' eggs on both Neare (X2c = 5.67, p < 0.05) and Farr (x2, = 69.3, p < 0.001) islands. At Port Colborne, 'early' eggs hatched significantly better than 'middle' (xzc = 24.04,~ < 0.001) and 'late' (xzc = 70.36, p < 0.001) eggs and 'middle' eggs hatched significantly better than 'late' eggs (xzc = 16.33, p < 0.001). Further, 'late' eggs at Port Colborne hatched significantly better than 'late' eggs at Neare (x2, = 13.81, p < 0.001) and Farr (x2, = 44.46, p < 0.001) Islands. The relationship between clutch size and time of clutch initiation to nest success at the Port Colborne colony is in Table 5. Within each timeperiod grouping, the number of eggs hatched per egg laid increased with clutch size and in two groupings the difference in hatching success was statistically significant (Table 6). Further, threeegg clutches showed a significant decline in hatching success with a delay in nest start time, whereas in two-egg clutches, a significant decline in hatching success was present in only one comparison (Table 6). The numbers of eggs that failed to hatch exceeded the numbers that hatched at the Neare

4 MORRIS ET AL. TABLE 2. Temporal distribution of egg laying and hatching of the common tern colonies. Complete data for Muggs I are unavailable because of water erosion of the colony Colony site Port Category Colborne Neare Farr Muggs I Muggs I1 First egg 4 May 8 May 8 May 6 May 4 June* 50% total egg production 25 May 22 May 22 May - 24 June 90% total egg production 6 July 8 July 3 July - 12 July First hatch 24 May 31 May 31 May - 25 June 50% total hatch 9 June 6 June 14 June - 13 July 90% total hatch 27 July 18 June 26 June - 22 July TABLE 3. Hatching success related to clutch size in the common tern colonies Clutch Clutches, Eggs hatched, Eggs hatched Mean hatch Colony size n n per egg laid per nest Port Colborne Neare Total Total Farr Total Muggs I Total Muggs I1 Total Island (72.3%), Farr Island (62.4%), and Muggs I1 (61.4%) colonies, whereas egg loss was lower at the Muggs I (29.6%) and Port Colborne (19.5%) colonies (Table 3). Categories that represent egg failure or loss are in Table 7. The most common category of egg loss was disappearance from the nest. The remaining categories varied among colonies in the extent to which they contributed to a reduced hatching success. Fledging Success With few exceptions, the numbers of chicks

5 CAN. 3. ZOOL. VOL. 54, 1976 TABLE 4. Hatching success related to time of clutch initiation in the comkon tern colonies Colony Date Nests, Eggs hatched, Mean hatch site grouping n % total laid per nest Port 4-28 May (88.1) 2.38 Colborne 29 May - 3 July (76.6) July - 8 Aug (52.7) 1.17 Neare 8-25 May (40.5) May - 8 July 13 3 (10.7) 0.23 Farr 8-25 May (65.3) May - 18 July 31 6 (6.6) 0.19 L 0 60 ;40- Z Pwf Colborne 2, b-k-x. Age (days) -x 4- -.%** #*. --*-. 2- r l / i X\~;X-X Neare Age (days) \ X-x FIG. 1. Numbers of common tern chicks known dead (broken lines) and disappeared (solid lines) in each of 3-day time intervals between hatching and fledging. that disappeared exceeded the numbers known to have died (Fig. 1). Most deaths and disappearances occurred among chicks less than 7 days old; however, losses continued up to 21 days posthatch, defined in this paper as the fledging age for common terns. Where sufficient data were available, we assessed the relationship between clutch size and fledging success (Table 8). The probability of a chick fledging once hatched did not appear to be related to clutch size. The number of chicks I fledged per egg hatched was very similar in all clutch sizes at the large Port Colborne colony. Furthermore, the difference in fledging success of chicks from two-egg vs. three-egg clutches was not statistically significant either at the Muggs I1 (xzc = 0.71, p > 0.1) or at the Port Colborne (xzc = 0.53, p > 0.1) colonies. In other words, although egg loss before hatching was clutchsize dependent (Table 3), chick loss after hatching was not (Table 8), at least in the two colonies analyzed.

6 MORRIS ET AL. Total Reproductive Success The total reproductive successes of all the common tern colonies are in Table 9. It is clear that the Port Colborne colony was substantially more successful than the other colonies in both measures of reproductive success (i.e. hatching and fledging success). The Port Colborne colony experienced a hatching success significant1 y greater than the Neare Tsland (xzc = 92.82, p < 0.001), Farr Island (x2, = 145.1,~ < 0.001) Muggs Z Ix2, = 13.29, p < $, and Muggs II (xzc = p < 0,001) colonies. Similarly, the Port Colborne colony experienced n fledging success significantly greater than the Neare Island (x2, = 6,82, p < 0.001), Fan Island (x2= = 21-73, p < 0.00 I), and Muggs I1 (X2, = 22.05, p < 0.001) colonies. There were no statistically significant differences in either the prehatch or posthatch mortality losses among all test combinations of the Neare Island, Farr Island, and Muggs I1 colonies. Chlorinated Hydrocarbon Residues The residue contents of eggs collected from the colonies are in Table 10. As we were unable to collect eggs from Neare Island and Farr Island in 1972, residue data shown were from egss collected in 1971 by M. Gilbertson and analyzed according to Reynolds (1969). Few demographic data are available from either colony from 1971, although numbers of tern nests (Neare = 28; Farr = 77) were similar to numbers in 1972 (Table 3). Gilbertson (1974) showed that 1,l-dichloro-2,2-bis(p-chloropheny1)ethene (DDE) and polychlorinated biphenyl (PCB) contamination in eggs at these colonies increased with time during the 1971 breeding season. Thus, for this analysis on the Neare and Farr Island samples, eggs laid before 31 May 1971 were separated from eggs laid after 31 May 1971 to determine whether differential accumulation of toxicants occurred. Considerable variation in organochlorine residues is reflected at all colonies (Table 10). Mann- Whitney U test analysis (Siege1 1956) revealed no significant differences in the residue levels of DDE or PCB among all test combinations of the Neare Island, Farr Island, and Muggs I1 colonies. However, there were statistically significant differences between the Port Colborne sample and each of the remaining samples. The DDE residue level in the Port Colborne eggs was sig-

7 1856 CAN. J. ZOOL. TABLE 6. The influence of clutch size and time of clutch initiation on hatching success in the Port Colborne common tern colony Contingency table analysis Clutch size Time period xz= 3 egg vs. 2 egg Early Middle Late egg Early vs, middle Middle vs. late Early vs. late 4.92* 3 egg Early vs, middle Middle vs. late Early vs. late *p < tp < nificantly lower than the total season sample at Neare Island (U = 4, p < 0.05), Farr Island (U = 3, z = -4.08, p < 0.001), and Muggs Island (U = I, p = 0.004). Similarly, the PCB residue level in the Port Colborne eggs was significantly lower than the total season sample at Neare Island (U = 4, p < 0.05), Farr Island (U = 6, z = -2.67, p = 0.003), and Muggs Island (U = 2, p = 0.008). Discussion All common tern colonies were followed throughout the breeding season from the time of first egg production until fledging was complete. Despite their geographic separation and differences in colony size, there was a high degree of synchrony in nest starts, hatching, and fledging dates. In addition, at all colonies large clutches hatched better than small clutches and 'early' clutches hatched better than 'late' clutches. Despite these similarities, the data show that the Port Colborne colony was more successful than the other colonies in all measures of reproductive success. We now consider factors contributing to these differences. Demographic Factors The reproductive success of a tern colony can be substantially influenced by the age structure of the breeding population. Low hatchability, late-starting nests, and small clutch sizes have previously been suggested as characteristics of young or inexperienced common terns nesting for the first time (Austin and Austin 1956; Langham 1968; LeCroy and Collins 1972). We do not have a direct measure of age structure in any of our colonies. However, an indirect measure is available by assessing the relative clutch-size distribution and success of nests started at different times during the summer in each colony. Three-egg clutches hatched better than twoegg clutches in all colonies (Table 3). Similarly, 'early' eggs had a significantly higher hatching rate than 'late' eggs (Table 4). A low overall reproductive performance could result therefore from a predominance of one- or two-egg clutches started relatively late in the season. A crude measure of the influence of clutch-size distribution in the various colonies on hatching success is gained by plotting the proportion of 'successful' three-egg clutches against hatching success per egg laid (Fig. 2). While the correlations between the two variables fall below the required level of significance in all three tests (0.1 > p > 0.05), it is of interest that the colonies are ranked in the same order as that established previously for overall reproductive success (Table 9). Low hatching success may also be caused by inadequate incubation, which may result from several factors. These include age and experience of the parents (Austin 1940), fishing ease and fishing potential within close range of the colony Proportion of three-egg Clutches per Total Clutches FIG. 2. The relationship between the number of eggs hatched per egg laid and the proportion of three-egg clutches per total clutches at the five common tern colonies.

8 MORRIS ET AL. (Lemmetyinen 1973), behavioural abnormalities caused by a high toxicant chemical load, which has been suggested in other species (McEwan and Brown 1966; Cade et al. 1968), and a variety of other factors. An indirect measure of incubation attention is the mean length of incubation in days. When these values for each colony were plotted against hatching success a statistically significant correlation was revealed (p < 0.05) in two of the three tests (Fig. 3). By this measure the colonies again are ranked in the same order as before (Table 9, Fjg. 2). Another demographic factor contributing to the greater success of the Port Colborne colony is suggested by the work of Darling (1938), who proposed that reproductive synchrony and reproductive success are influenced by social stimulation, which increases with colony size. Our data do not suggest that the large Port Colborne colony was more tightly synchronized than the other smaller colonies (Table 2), although it was clearly more successful. Tentative evidence in support of the occurrence of a 'Darling effect' in common tern colonies was provided by Austin (1940, 1945) and Palmer (1941), although Haartman (1945) suggested that larger colonies studied by Darling may have contained a greater proportion of older more experienced birds, which in common terns (Cooper et al. 1970) and kittiwakes (Coulson and White 1958, 1960) are known to be more successful than first nesters. Nelson (1966) reviewed the evidence in support of the synchronization theory, and its adaptive advantages to seabird colonies. He suggested a relationship between colony size and breeding success from his work with gannets (Sula bassana) in Scotland but was unable to comment on factors mediating it. The social stimulation theory as a synchronizing force in nesting colonies bird species has also been questioned by Orians (1961), who proposed that it neglects important aspects of the environment which vary temporally and spatially. He allows, however, that "group stimulation may operate in nomadic flocks prior to the initiation of breeding" (Orians 1961, p. 337). Despire these qualifications, we believe that colony size may be an important indicator of reproductive success. We admit that demographic factors considered in this section are either indirect or speculative. However, it seems clear than an assessment of their contribution is essential when interpreting '? '? N - ~ ~ N W mat- I

9 1858 CAN. J. ZOOL. VOL. 54, 1976 TABLE 8. Fledging success related to clutch size in the Port Colborne and Muggs I1 common tern colonies Port Colborne Muggs I1 Chicks Mean Chicks Mean Eggs Chicks fledged fledged Eggs Chicks fledged fledged Clutch hatched, fledged, per egg per hatched, fledged, per egg Per size n n hatched nest n n hatched nest Totals TABLE 9. The total reproductive performance of the common tern colonies Eggs Chicks Chicks hatched Mean fledged fledged Eggs per egg hatch Chicks per eggs Per egg Mean hatched, laid, per fledged, hatched, laid, fledged Colony n +1 SE nest n +1 SE + 1 SE per nest Port Colborne k0.008 Neare Island k0.043 Farr Island k0.037 Muggs I k0.020 Muggs I k0.029 TABLE 10. Residue content of eggs collected from the common tern colonies Residue level, ppm* -... PCB p,pj-dde DDE: PCB ratio Collection Arithmetic Geometric Arithmetic Geometric (geometric Colony dates n mean mean mean mean mean) Port Colborne 17 June Aug ( )t ( ) May, ( ) ( ) Neare 17 May Island 12 July, 1971 ( ) ( ) May, ( ) ( ) Farr 17 May Island 5 July, 1971 ( ) ( ) Muggs I1 5 June July, 1972 ( ) ( ) "Expressed on dry matter basis. trame.

10 MORRIS ET AL Farr Mean Incubation Time (Days) FIG. 3. The relationship between the number of eggs hatched per egg laid and 'parent attentiveness' measured as mean incubation time at the five common tern colonies. the influence of environmental factors on the reproductive success of a common tern colony. Predation Paynter (1949) commented on predation of common tern eggs and chicks by gull species and Hatch (1970) suggested that the annual toll of common tern chicks taken by herring gulls may be as high as chicks per adult tern pair. Herring gulls and ring-billed gulls were nesting close to several tern colonies in our study and may have been responsible for the disappearance of both eggs (Table 7) and chicks (Fig. 1). We never saw common tern eggs taken by either gull species and only a few chicks were attacked. Leck (1971) reported feeding by ring-billed gulls at night; however, our nocturnal visits to the colonies never revealed evidence of predation. We believe it unlikely therefore that gull predation was a major factor contributing to differences in reproductive success among the colonies. Nocturnal predation was identified in 1973 at the Port Colborne colony when on several occasions one of us (R.A.H.) observed a blackcrowned night heron (Nycticorax nycticorax) feeding on tern eggs and chicks (Hunter and Morris 1976). The contribution to egg and chick loss of such predation in 1972 is unknown. Nest Site Competition by Gulls Proximity of large colonies of ring-billed gulls may have a major influence on the long-term stability and reproductive success of common tern colonies (Morris and Hunter 1976~). Ringbilled gulls arrived and began nesting about 2 weeks in advance of common terns at the Muggs and Port Colborne colonies and at Muggs Island, occupied nesting sites used by terns in previous years. Intensive observations throughout the breeding season provided no evidence that gull interference at the Muggs Island colony resulted in the reduced reproductive success of the terns. Food Availability Evans and McNicholl (1972) proposed that the amount of food available during or immediately before laying was an important proximate factor influencing clutch size in an Arctic tern (S. yaradisaea) population along the west coast of Hudson Bay, Canada. Lemmetyinen (1972) suggested that increased mortality and a sudden reduction of growth among 1- to 6-dayold Arctic tern chicks near Spitsbergen, Norway, were related to a shortage of fish food or to a reduction in fishing potential. He proposed that higher populations of fish close to the shore resulted in larger clutches and earlier laying dates in Arctic tern colonies on inshore compared with offshore islands (Lemmetyinen 1973). Nisbet (1973) showed that fledging success in common terns was correlated with egg size and with the performance of the male during 'courtship feeding.' He also suggested that the amount of food brought by the male while the female was brooding the chicks was an important indication of the future fledging success of the brood. We did not directly measure the potential food supply at any of the common tern colonies. However, both 'fishing potential' and 'fishing ease' were considerably better at Port Colborne than at the other colonies. In May and June 1972, terns from the Port Colborne colony were frequently seen feeding on minnow schools within 500 m of the colony site. These schools were sufficiently abundant to create a 'false bottom' effect. No similar proximate feeding source was ever observed near any of the other colonies. These observations may be reflected in the higher overall reproductive success of the Port Colborne colony. Adults that have little difficulty feeding

11 1860 CAN. J. ZOOL. VOL. 54, 1976 themselves and their chicks would spend more time incubating and raise chicks more successfully than parents that are forced to forage widely for food, leaving eggs and chicks unattended. Floodinn u Loss of common tern eggs and chicks by flooding was recorded by Austin (1940), Rooth (1958) Ludwig (1962), and Nisbet (1972). Farr Island and Neare Island experienced moderate egg loss owing to flooding (Table 7) and the Muggs I colony was severely disrupted because of erosion bv water. All these colonies were more susceptible to water damage than the more elevated Port Colborne colony. We therefore recognize flooding as a factor influencing the reproductive success of common terns, although the extent of egg loss directly attributed to it was less than several other categories (Table 7). Toxicant Chemical Influence An inverse relationship between organochlorine residues in the eggs and tissues of colonial ground-nesting birds and reproductive success has been inferred in a number of studies (Gilbertson and Reynolds 1972; Hickey and Anderson 1968; Keith 1966; Koeman et al. 1968). The marked difference in reproductive success between the Port Colborne and the other colonies may therefore be related to the lower toxicant load in eggs (and presumably adult individuals) at the former site (Table 10). However, several lines of evidence suggest a need for some caution. First, the egg-shell thickness index (Ratcliffe 1967) calculated for eggs taken from Neare Island in 1971 (n = 38) showed a 2.9% decrease from National Museum samples collected between 1886 and Eggs collected from Port Colborne in 1971 (n = 9) showed a corresponding decrease of 7.9% (Gilbertson, unpublished data). These findings contradict the inverse relationship expected between toxicant load and egg-shell thickness. Second, if toxic chemicals alone were responsible for reducing reproductive success in common tern colonies, adjacent colonies started at the same time by adults foraging in the same local area should experience similar reproductive success (or lack of it). Neare Island and Farr Island were separated by 500 m of open water and common terns began nesting on both sites on the same day in 1972 (Table 2). The hatching rate of eggs on Farr Island, while short of the required level of significance (x2, = > p > 0.05) was greater than that of eggs on Neare Island (Table 3). Also, the mean number of chicks fledged per nest on Farr Island was greater than twice the number fledged per nest on Neare Island (Table 9). Residue levels of DDE and PCB in eggs from the two colonies were very similar (Table 10) with no significant differences (Mann-Whitney U test analysis) between the May 1971 samples or the total season samples. Thus, the reduced reproductive success of the tern colony on Neare Island may have been related more directly to differences in the suitability of nesting habitat, to differences in the structural composition of the two populations, or to other undefined factors rather than to toxicant chemical load. At present we have insufficient data to adequately assess these alternatives. In studies of the reproductive success of colonial seabirds we recognize the importance of selecting sample areas from larger colonies that are fully representative of the total colony and that include early as well as late nesters. This is of particular importance when reproductive success is compared among colonies at widely separated geographic locations where data are collected by different groups of investigators often using different methods. We note, for example, that the highest fledging success ever reported for common terns (2.09 chicks per nest) was from a 35-nest fenced section of a larger colony on Bird Island, Massachusetts (Nisbet and Drury 1972). In this instance the authors correctly point out that while such a fenced sample may be representative of birds nesting at the peak period of the season, it may not provide complete information about the performance of the total colony. Finally, we note the substantial proportion of eggs and chicks lost at all our colonies which could not be accounted for (i.e. disappeared) despite fencing of the colonies and regular visits to them. This observation causes us some concern when viewed in the context of other published studies on the breeding biology of common terns. Langham (1968) visited his common tern colonies daily, which to our knowledge represents the most intensive long-term visitation program yet reported. We recognize the difficulties inherent in studying large isolated colonies but,suggest that visits at weekly or longer intervals

12 MORRIS ET AL (Cooper et al. 1970; Hays 1970) may seriously jeopardize the validity of data collected and their interpretation. In conclusion our data demonstrate clearly the substantial variations that occur in the reproductive success of common tern colonies. Frequently, no single factor can be completely associated with observed differences and we suggest that considerable caution should be exercised when considering the relationships between reproductive success and factors influencing it. Acknowledgements We gratefully acknowledge the generous financial assistance of Environment Canada, Canadian Wildlife Service, Toxic Chemicals Section, and the National Research Council of Canada (operating grant No. A6298 to R.D.M.). We thank the following individuals and organizations, who provided a variety of services and advice during this and related studies: J. Bonisteele, Canada Centre for Inland Waters, D. Christie, W. Corkum, R. Fabro (Algoma Steel Co., Canada Furnace Division), M. Gilbertson, J. Keith, A. Kendrick, E. Muller, R. Nelder, S. Powell, R. Profit Sr., R. Profit Jr., H. Van der Veen (Toronto Island Sailing School), and the Ontario Hydro Commission. AUSTIN, 0. L., Some aspects of the individual distribution in the Cape Cod tern colonies. Bird-Banding, 13: The life span of the common tern. Bird- Banding, 13: The role of longevity in successful breeding of the common tern. Bird-Banding, 16: AUSTIN, 0. L., and 0. L. AUSTIN, JR Some demographic aspects of the Cape Cod population of common terns (Sterna hirundo). Bird-Banding, 27: CADE, T. J., C. M. WHITE, and J. R. HAUGH Peregrines and pesticides in Alaska. Condor, 70: COOPER, D., H. HAYS, and C. PESSINO Breeding of the common and roseate terns on Great Gull Island. Proc. Linn. Soc. N.Y. 71: COULSON, J. C., and E. WHITE The effect of age on the breeding biology of the kittiwake, Rissa tridactyla. Ibis., 100: -~ 651. ~ The effect of age and density of breeding birds on the time of breeding in the kittiwake, Rissa tridactyla. Ibis, 102: DARLING, F. F Birds flocks and the breeding cycle. University Press, Cambridge, England. EVANS, R. M., and M. K. MCNICHOLL Variations in the reproductive activities of Arctic terns at Churchill, Manitoba. Arctic, 25: GILBERTSON, M Seasonal changes in organo- chlorine compounds and mercury in common terns of Hamilton Harbor, Ontario. Bull. Environ. Contamin. Toxicol. 12: GILBERTSON, M., and L. M. REYNOLDS Hexachlorobenzene in the eggs of common terns in Hamilton Harbor, Ontario. Bull. Environ. Contamin. Toxicol. 7: GILBERTSON, M., R. D. MORRIS, and R. A. HUNTER Abnormal chicks of colonial fish-feeding birds on the lower Great Lakes. Auk. In press. HAARTMAN, L. VON Zur biologie der Wasser-und infervogel in Schanenmeer Sudwest-Finnlands. Acta Zool. Fenn. 44: HATCH, J. J Predation and piracy by gulls at a ternery in Maine. Auk, 87: HAYS, H Great Gull Island report on nesting species Proc. Linn. Soc. N.Y. 71: HICKEY, J. J., and D. W. ANDERSON Chlorinated hydrocarbons and eggshell changes in raptorial and fish-eating birds. Science, 162: HOWARD, D. V Do New England terns have a future? Mass. Audubon, September, HUNTER, R. A., and R. D. MORRIS Nocturnalpredation by a black-crowned night heron at a common tern colony. Auk. 93: KEITH, J. A Reproduction in a population of herring gulls (Larus argentatus) contaminated by DDT. J. Appl. Ecol. (Suppl.) 3: KOEMAN, J. H., J. VEEN, E. BROUWER, L. HUISMAN- DE-BROUWER, and J. L. KOOLEN Residues of chlorinated hydrocarbon insecticides in the North Sea environment. Helgol. Wiss. Meersunters. 17: LANGHAM, N. P The biology of terns, Sterna spp. Ph.D. Thesis, University of Durham, England. LECK, C. F Nocturnal habits of ring-billed gulls (Larus delawarensis) at Thimble Shoal, Virginia. Chesapeake Sci. 12: LECROY, M., and C. T. COLLINS Growth and survival of roseate and common tern chicks. Auk, 89: LEMMETYINEN, R Growth and mortality in the chicks of Arctic terns in the Kingsfjord area, Spitsbergen, in Ornis Fenn. 49: Clutch size and timing of breeding in the Arctic Tern in the Finnish archipelago. Ornis Fenn. 50: LUDWIG, J. P A survey of the gull and tern populations of Lakes Huron, Michigan and Superior. Jack-Pine Warbler, 40: Herring and ring-billed gull populations of the Great Lakes. Publ. 15. Great Lakes Res. Div. Nos Univ. Mich. Publ. MARSHALL, N Night desertion by nesting common terns. Wilson Bull. 34: MCEWEN, L. C., and R. L. BROWN Acute toxicity of dieldrin and malathion to wild sharp-tailed grouse. J. Wildl. Manage. 30: MORRIS, R. D., and R. A. HUNTER. 1976a. Factors influencing desertion of colony sites by common terns (Sterna hirundo). Can. Field-Nat. 90: Monitoring incubation attentiveness of ground-nesting colonial seabirds. J. Wildl. Manage. 40: NELSON, J. B The breeding biology of the Gannet

13 1862 CAN. J. ZOOL. VOL (Sula bassana) on the Bass Rock, Scotland. Ibis, 108: NISBET, I. C. T Disaster year for terns. Man Nature, Dec Courtship-feeding, egg-size and breeding success in common tems. Nature, 241: N~SBET, I. C. T., and W. H. DRURY Measuring breeding success in common and roseate tems. Bird- Banding, 43: ORIANS, G. H Social stimulation within blackbird colonies. Condor, 63: PALMER, R. S A behavioral study of the common tern. Roc. Boston Soc. Nat. Hist. 42: PAYNTER, R. A Clutch size and egg mortality of Kent Island herring gulls. Ecology, 30: RATCLIFFE, D. A Decrease in egg-shell weight in certain birds of prey. Nature, 215: REYNOLDS, L. M Polychlorinated biphenyls (PCB's) and their interference with pesticide residue analysis. Bull. Environ. Contam. Toxicol. 4: ROOTH, J Relations between black-headed gulls (Larus ridibundus) and terns (Sterna spp) in the Netherlands. Bull. Int. Comm. Bird Preserv. 7: SIEGEL, S Non-parametric statistics. McGraw-Hill, New York.