We investigated nest predation in a population of Florida Scrub Jays (Aphelocoma c. coerulescens) at Archbold Biological Station, Lake Placid, Florida. Frequency of nest visits by investigators did not influence predation rates. Nest predation increased as the breeding season progressed and occurred most often during daylight hours. When seasonal effects were held constant, nestlings were depredated more often than eggs, and young nestlings were depredated more often than old nestlings. Several lines of evidence indicated that, for Florida Scrub Jays, diurnal snakes and birds were the most important nest predators, while nocturnal mammals were relatively less important. Late in the breeding season, pairs with helpers experienced less nest predation than pairs without helpers. This effect was primarily the result of reduced predation on nestlings. Received 17 June 1991, accepted 13 January 1992.

The Auk 109(3):585-593, 1992 Department of Biology, University of South Florida, Tampa, Florida 33620, USA; and 2Department of Biology, Allegheny College, Meadville, Pennsylvania 16335, USA NEST predation is a primary source of egg and nestling mortality in many species (Martin in press), and is increasingly recognized as a sig- nificant process shaping avian life-history char- acteristics, habitat selection, and community structure (e.g. Slagsvoid 1982, Martin 1988a, b, in press). Patterns of nest predation also may interact with habitat fragmentation in ways that have important implications for conservation biology (e.g. Loiselle and Hoppes 1983, Wilcove 1985). However, despite the growing recogni- tion of the importance of nest predation, our understanding of the ecological factors that in- fluence the behavior of specific nest predators and rates of nest predation in natural popula- tions remains limited (Martin 1987, in press). In this paper we examine the ecological and social factors that may affect predation on eggs and nestlings of the Florida Scrub Jay (Aphelo- coma c. coerulescens). This jay is restricted to pen- insular Florida where it inhabits recently burned oak scrub (Woolfenden and Fitzpatrick 1984). Nests are typically positioned about 1 m above the ground in oak shrubs (Woolfenden 1974). Nesting is from March through June. Most clutches are of three or four eggs, rarely two or five. Incubation begins after the last or penul- timate egg is laid. Eggs hatch after approxi- mately 18 days of incubation, and nestlings fledge about 18 days after hatching (Woolfen- den and Fitzpatrick 1984). 3present address: Bionetics Corporation, B10-2, Kennedy Space Center, Florida 32899, USA. Florida Scrub Jays exhibit a cooperative breeding system in which about one-half of all breeding pairs are assisted by nonbreeding adults (helpers). For pairs with helpers the modal number of helpers is 1, the mean is 2, and the maximum is 6 (Woolfenden and Fitzpatrick 1990). Helpers assist in feeding young, defend- ing the territory, guarding the nest and mob- bing predators. They do not build nests, incu- bate, or brood. Although breeding pairs assisted by two or more helpers are no more successful than pairs with only a single helper, unassisted pairs produce significantly fewer offspring than pairs with helpers (Woolfenden and Fitzpatrick 1984, Mumme in press). The major way that helpers appear to increase reproductive success is by reducing predation on eggs, nestlings (Woolfenden 1978) and fledglings (McGowan and Woolfenden 1989). Helpers could reduce diurnal predation by serving as sentinels (McGowan and Woolfenden 1989), or by mob- bing potential predators (Francis et al. 1989), or both. However, helpers would be ineffective at reducing the frequency of nocturnal predation (Woolfenden and Fitzpatrick 1984). Based on data collected over a 10-year period (1969-1979), Woo lfenden and Fitzpatrick (1984) concluded that predation is the primary cause of nest failure in the Florida Scrub Jay, account- ing for 67% of egg loss and 85% of nestling loss. Thus, an evaluation of factors influencing nest predation is critical for a thorough understand- ing of the breeding biology of this species. In this paper, we provide such an evaluation by focusing on the following questions: (1) Does the frequency at which investigators visit nests influence predation rates? (2) What are the ef- fects of habitat and proximity to vehicle trails? (3) How is the frequency of nest predation in- fluenced by season and nest stage? (4) Does nest predation in the Florida Scrub Jay occur pri- marily during the day or the night? (5) Which predators are primarily responsible? (6) How does the presence of nonbreeding helpers affect the rate of nest predation? In addressing these questions, our study extends earlier analyses (e.g. Woolfenden and Fitzpatrick 1984) and pro- vides several revised interpretations of previ- ously published work on this species. METHODS We worked on marked jays that reside on the prop- erty of the Archbold Biological Station (ABS), a 2,000- ha research station located 12 km south of Lake Placid, Highlands County, Florida. Since 1969, Woolfenden and colleagues have monitored a subset of the total population of Florida Scrub Jays resident at ABS. Up to 50 groups of Scrub Jays are censused every month and all nests are found each year. Almost all nests are found before or very early in incubation. Through these efforts the age, sex, and breeding histories of many individuals are known. In 1987, Schaub and Mumme conducted detailed observations on 76 nests with contents owned by 49 groups located in open oak scrub immediately south of the groups monitored by Woolfenden and col- leagues. Most nests (n = 60) were found before in- cubation began, and all nests were monitored until they either fledged young or failed. Schaub visited some nests twice daily (n = 41) at sunrise (+30 min) and sunset (+ 30 min) and others approximately every third day (n = 35). Acts of predation that occurred between the sunrise and sunset nest checks are con- sidered diurnal, and those that occurred between the sunset and sunrise nest checks are considered noc- turnal. To determine seasonal and circadian activity rates of the snakes and mammals that are the probable nest predators, Schaub censused vertebrate tracks during most of the 1987 nesting season. Tracks on a strip of sand 1.1 km long and 1 m wide, which extended along a broad firelane that coursed through the study site, were censused twice daily immediately after the sun- rise nest checks and immediately before the sunset nest checks. Tracks made between the sunrise and sunset censuses are considered diurnal, and those made between the sunset and sunrise censuses are consid- ered nocturnal. Terrestrial-predator activity is calcu- lated as the number of tracks per number of hours since the last census. For example, a census showing tracks of five potential predators, which were known to have been made during the preceding 10 h, results in an activity index value of 0.5. If rain obscured the tracks (21 of 100 mornings and 11 of 91 evenings), no census was made. Sightings of potential avian nest predators also were recorded and used to determine their activity rates. The rate of avian predator activity was calculated as number of visual sightings per hour of field time within the study tract. The activity rates of terrestrial and avian predators were combined into a single index, which was calculated as the number of observations (tracks plus sightings of birds) per observation hour (tracking hours plus field hours). Data from 552 nesting attempts from the years 1974- 1979 and 1981-1987 combined were used to calculate predation rates. The years 1969-1973 were excluded because historical information about the breeders was sparse, and the year 1980 was excluded because of the social disruption caused by a major die-off in fall 1979 (Woolfenden and Fitzpatrick 1991). The number of fledglings produced per group within a breeding sea- son was used to make between-year comparisons. For this measurement of nesting success the years 1974- 1987, excluding 1980, were not significantly different (Kruskal-Wallis, H = 17.33, df = 12, P = 0.14). We deleted from our analyses the few nests that failed to reach the incubation stage, or that experienced loss or injury of a breeder. Daily predation rates were calculated as the number of apparent acts of predation per number of days a nest contained either eggs or nestlings. Losses that occurred at an unknown time between nest visits were considered as having occurred midway between vis- its. Calculation of daily rates of predation are based on three different ways of recognizing apparent nest pre- dation: (1) Individual acts--losses of all or part of a clutch or brood between consecutive nest visits that cannot be attributed to a cause other than predation. (2) Ultimate failures--nests that eventually lose all eggs or nestlings as a result of one or more acts of apparent predation. (3) Instantaneous failures--nests that fail as a result of loss of all eggs or nestlings between con- secufive nest inspections. Predation rates derived from rates of individual acts of predation thus provide the most liberal estimate of nest predation, while esti- mates derived from the rate of instantaneous failure provide the most conservative estimate. For Florida Scrub Jays, we think the most liberal method provides the most accurate measure. Starvation of nestlings and subsequent brood reduction is a rare event for Florida Scrub Jays (Woolfenden and Fitzpatrick 1984). In the few cases observed during 1987, one member of a brood became visibly smaller then its siblings early in the nestling period, usually by day 3. Most of these runts survived to day 12. Therefore, we sus- pect that brood reduction that is the direct result of starvation rarely goes undetected in Florida Scrub Jays. The existence of long-enduring runts suggests that most instances of partial brood loss do in fact represent instances of partial predation, which sup- T^BEE I. Effect of habitat type on rates of nest predation, as measured by both daily rates of individual acts and overall ultimate nest failures, 1974-1979 and 1981-1987 combined. Habitat type Open scrub Overgrown scrub Pasture Total nest-days Individual acts of predation Predation rate G = 2.25, n = 16,037, P > 0.25 Nests that produced young Nests failing because of predation Predation rate G = 7.59, n = 463, P < 0.025 13,193 2,113 731 244 44 19 0.0185 0.0208 0.0260 258 40 12 112 27 14 0.3027 0.4030 0.5385 ports our incorporating partial losses into estimates of nest predation rates. The time period within the nesting season, expe- rience of the breeders, and age of the breeders all affect nesting success in Florida Scrub Jays (Woolfen- den and Fitzpatrick 1984). To control for the effect of season, we divided the nesting season into half-month (15-16 day) and month (30-31 day) intervals. To re- duce the effects of breeder inexperience and senes- cence, in analyses of helper contributions we deleted all first breeders and those few breeders older than 11 years. Because sample sizes and variances were unequal, nonparametric statistics were used to analyze daily rates of predation and predator activity. Categorical data were analyzed via contingency tables using log- likelihood ratio tests (G-test) with Williams' correc- tion (Sokal and Rohlf 1981), or chi-square tests with Yates' correction for continuity (Zar 1984). Some of the following analyses use individual nests rather than nest days as units in order to more closely meet the assumption of independence. RESULTS Effects of investigator visitation.---Rates of nest predation in 1987 were not significantly influ- enced by the two different nest-visitation treat- ments. Nests visited by investigators twice daily experienced 32 individual acts of apparent pre- dation in 1,155 nest-days (0.028 acts/day), com- pared to 22 individual acts in 979 nest-days (0.022 acts/day) for nests visited approximately every three days (G = 0.37, P > 0.9). Similarly, 19 (46.3%) of the 41 nests visited twice daily ulti- mately failed due to predation, compared to 17 (48.6%) of the 35 nests visited every three days (G = 0.38, P > 0.9). Effects of habitat and proximity to vehicle trails.- In examining the effects of habitat on the rate of nest predation, three major habitats were ex- amined: recently burned open scrub (burned within last 20 years), unburned overgrown scrub (not burned for more than 20 years), and cattle pasture with scattered oak shrubs, palmetto clumps and tall pines. Although the daily rate of nest predation (individual acts) is similar for all three habitats, the rate at which nests ulti- mately failed because of predation differs sig- nificantly among habitats, with open scrub ex- hibiting the lowest rate (Table 1). We analyzed the effects of proximity to ve- hicle trails on nest predation by dividing nests into two categories: those within 15 m of a ve- hicle trail, and those farther away. We consider it unlikely that a potential predator searching for prey from these trails would detect a nest beyond 15 m into the scrub. Analysis shows proximity to trails had no effect on nest pre- dation. Nests within 15 m of vehicle trails ex- perienced 104 individual acts of predation in 6,137 nest-days (0.017 acts/day), compared to 183 individual acts in 9,372 nest-days (0.019 acts/ day; G = 1.23, P > 0.25). Over the entire breed- ing cycle 54 of 172 nests (31.4%) within 15 m of a trail ultimately failed because of predation, compared to a similar 85 of 271 (31.4%) of the more distant nests (G = 0.00, P > 0.9). Effects of season, nest stage, and time of day.- All three methods of calculating daily rates in- dicate that nest predation increased as the sea- son progressed (Fig. 1). Daily rates of predation, calculated by the number of individual acts, the number of ultimate failures, and the number of instantaneous failures were all significantly correlated with the advance of the breeding sea- son (rs = 0.96, 1.00, and 0.96, respectively; n = 7, P < 0.05 for all). Predators took relatively more nestlings than eggs (Fig. 2). The overall daily rate of egg pre- 0.05 LD <: 0.04 Z . 0.03 n' 0.02 O.Ol o.oo [] INDIVIDUAL ACTS [] ULTIMATE FAILURES [] iNSTANTANEOUS FAILURES 1 2 3 4 5 6 7 PERIOD WITHIN NESTING SEASON Daily predation rates on nests of Florida Fig. 1. Scrub Jays calculated three different ways (1974-1979 and 1981-1987 combined). Half-month periods with- in nesting season are: (1) 9-24 March, (2) 25 March- 8 April, (3) 9-23 April, (4) 24 April-8 May, (5) 9-23 May, (6) 24 May-7 June, and (7) 8-23 June. dation of 0.0072 (66 individual acts in 9,204 nest days) was significantly lower than the overall daily rate of nestling predation of 0.0285 (212 individual acts in 7,428 nest days; G = 117.06, P < 0.001). Furthermore, the rate of predation on nestlings was higher than that on eggs dur- ing six of the seven half-month periods of the nesting season (Wilcoxon T = 1, n = 7, P = 0.028). Similar results were obtained in analyses controlling for age and experience of breeders (Schaub 1990). Young nestlings (day 8 or younger) were taken by predators more often than older nestlings 0.06 I.-- 0.05 Z 0.04 ' 0.03 Q_ 0.02 : o.ol 0.00 [] EGGS [] NESTLINGS 1 2 3 4 5 6 PERIOD WITHIN NESTING SEASON Fig. 2. Daily predation rates (individual acts) on eggs and nestlings of Florida Scrub Jays (1974-1979 and 1981-1987 combined). Asterisks indicate a log- likelihood ratio P < 0.05. Half-month periods within nesting season as in Figure 1. 0.20 r 0.15 0.10 0.05 0.00 [] YOUNG NESTLINGS [] OLD NESTLINGS DATA , ..... 1 2 3 4 5 6 7 PERIOD WITHIN NESTING SEASON Fig. 3. Daily predation rates (individual acts) on young (day 8 or younger) and old (day 9 or older) nestlings of Florida Scrub Jays (1974-1979 and 1981- 1987 combined). Asterisks indicate a log-likelihood ratio P < 0.05. Half-month periods within nesting season as in Figure 1. (Fig. 3). The overall daily rate of predation on the younger nestlings (0.0404, 115 individual acts in 2,847 nest days) was significantly greater than for older nestlings (0.0214, 40 individual acts in 1,869 nest days; G = 13.53, P < 0.001). Predation rates on younger nestlings consis- tently exceeded that on older nestlings during all six half-month periods for which data were available (Wilcoxon T = 0, n = 6, P = 0.032) and were significant for two periods (log-likelihood ratio P < 0.05). Dawn and dusk nest checks allowed 32 in- dividual acts of predation at 41 nests to be cat- egorized as either diurnal (n = 23) or nocturnal (n = 9). Thus, diurnal predation occurred more than twice as often as nocturnal predation, and the difference is significant (X 2 = 5.28, df = 1, P < 0.025). Because some acts of predation cat- egorized as nocturnal may have occurred in the dim light shortly before the early morning nest checks or after the evening nest checks, the true frequency of diurnal predation may have been even higher. Activity of potential nest predators.--Track cen- suses revealed that most of the potential mam- mal nest predators were active at night, and most of the potential snake nest predators were active during the day. In 1987, 1,814 of 1,846 (98.3%) mammal tracks were made between the evening and morning track censuses, compared to just 31 of 544 (5.7%) snake tracks (G = 1,996, P < 0.0001). Furthermore, some of the few snake tracks counted during the morning censuses may TAI3L� 2. Rates of individual acts of predation (1974-1979 and 1981-1987 combined) and rates of potential nest-predator activity (1987). Predator activity Diurnal Acts of All Nocturnal Diurnal Diurnal snakes and Half-month periods predation" predators mammals birds snakes birds 9-24 March 2/870 259/565 0.0023 0.46 25 March-8 April 30/3,556 280/555 0.0084 0.50 9-23 April 85/4,951 630/584 0.0172 1.08 24 April-8 May 87/3,587 458/559 0.0243 0.82 9-23 May 48/2,067 339/418 0.0232 0.81 24 May-7 June 44/1,312 474/434 0.0335 1.09 8-23 June 21/439 242/245 0.0478 0.99 rs with period in season 0.68 P > 0.05 rs with predation rate 0.71 P = 0.05 196/263 0.75 175 213 0.82 464 202 2.30 292 194 1.51 222 141 1.57 309 152 2.03 136/81 1.68 0.61 > 0.05 0.57 > 0.10 32/223 23/79 55/302 0.15 0.29 0.18 31/225 67/117 98/342 0.14 0.57 0.29 62/224 95/158 157/382 0.28 0.60 0.41 76/200 84/163 160/363 0.38 0.52 0.44 35/176 80/101 115/277 0.20 0.79 0.42 36/132 93/150 129/282 0.27 0.62 0.46 20/78 71/85 91/163 0.26 0.84 0.56 0.39 0.86 0.96 > 0.10 < 0.025 = 0.0025 0.54 0.71 1.00 > 0.10 = 0.05 <0.005 Individual acts of predation per number of nest-days. For mammals and snakes tracks/tracking hour; for birda sightings/field hour, and rates for both. have been made during the brief periods of light before the morning censuses or, more like- ly, after the evening censuses. Most of the mam- mal tracks (>80%) were made by raccoons (Pro- cyon lotor). When all track counts and avian predator sightings were combined to form an index of the activity of all potentially important nest predators, no significant correlation exists with the advance of the nesting season, but one does exist with the rate of nest predation as measured with individual acts (Table 2). Neither an index of nocturnal mammal predator activity nor an index of diurnal avian predator activity is sig- nificantly correlated with the advance of the nesting season or the rate of individual acts of nest predation. Diurnal snake activity, how- ever, is significantly and positively correlated with both the advance of the nesting season and with the rate of nest predation (individual acts). Because most nest predation is diurnal, the activities of diurnal snakes and birds are combined into a single index of potential di- urnal predator activity. The activity rate of di- urnal snakes and birds was significantly cor- related with the advance of the nesting season and was perfectly correlated with the rate of individual acts of nest predation (Fig. 4). Effects of helpers.--In general, the presence of helpers had little effect on rates of individual acts of nest predation (Fig. 5). However, during the final month of the breeding season, pairs without helpers experienced significantly high- er rates of nest predation than pairs with help- ers (Fig. 5A). This apparent effect of helpers on late-season nest predation was attributable to a significantly reduced rate of predation on nests with nestlings, but not nests with eggs (Figs. 5B and 5C). I-- z ill ._l 0.06 0.05 0.04 0.03 0.02 0.01 0.00 0.1 r s = 1.00 P: 0.001 0.3 0.4 0.5 0.6 0.7 INDEX OF DIURNAL PREDATOR ACTIVITY Fig. 4. Relationship between an index of diurnal predator activity and daily predation rates (individual acts) during seven half-month periods of nesting sea- son shown in Table 2. A. NESTS WITH EGGS OR NESTLINGS 0.12 , 0.10 0.08 0.06 0.04 >'; 0.02 0.00 [] HELPERS PRESENT [] HELPERS ABSENT MARCH APRIL MAY JUNE B. NESTS WITH EGGS 0.12 I.U , 0.10 z_ 0.08 ' 0.06 IJJ 0.04 0.02 0.00 [] HELPERS PRESENT [] HELPERS ABSENT MARCH APRIL MAY JUNE LU C. NESTS WITH NESTLINGS 0.12 [] HELPERS PRESENT 0.10 [] HELPERS ABSENT 0.08 0.06 0.04 0.02 NO DATA 0.00 MARCH APRIL MAY JUNE MONTH Fig. 5. Daily predation rates (individual acts) on nests with (A) eggs or nestlings, (B) eggs, and (C) nestlings for Florida Scrub Jay pairs with helpers present and helpers absent (1974-1979 and 1981-1987 combined). Nests belonging to novice and senescent breeders excluded from analysis. Asterisks indicate a log-likelihood ratio P < 0.05. DISCUSSION Identifying the predators.--The eggs and nest- lings of Florida Scrub Jays are potential food for many possible predators. The continuing 22- year study of Scrub Jays at Archbold Biological Station (Woolfenden and Fitzpatrick 1990) has implicated several species. One snake, the east- ern coachwhip (Masticophis fiagellum) is known to take nestlings (Westcott 1970), and it and the eastern indigo snake (Drymarchon corais) are known to take fledglings (Webber 1980, Mumme 1987). Convincing evidence of nest predation, including direct observations, also exist for the Red-tailed Hawk (Buteo jamaicensis), Eastern Screech-Owl (Otus asio), Great Horned Owl (Bubo virginianus), and bobcat (Lynx rufus), as well as for the Northern Harrier (Circus cyaneus), a northern migrant which may be present through late spring. Despite the enormous number of hours of field observations, the relative impor- tance of these and other potential Scrub Jay nest predators (e.g. Swallow-tailed Kites, Elanoides forficatus; Fish Crows, Corvus ossifragus; Blue Jays, Cyanocitta cristata; and raccoons) has been un- clear (Schaub 1990). Our study, however, has produced several lines of evidence indicating that mammalian predators are relatively less important than are snakes and birds. First, twice-daily nest checks and track censuses indicate that, although at least two-thirds of the predation on Florida Scrub Jay nests occurs during daylight hours, poten- tial mammalian predators are overwhelmingly nocturnal in their activity patterns. Second, re- suits from radio-tracking studies conducted at ABS indicate that potential mammalian preda- tots such as bobcats and raccoons commonly use vehicle trails as avenues for travel (Worley 1980, Wassmer et al. 1988). If bobcats and raccoons were important as predators upon jay nests, proximity of nests to the trails might correlate with increased nest predation (Best 1978, Kep- pie and Herzog 1978). However, we found that the rate of Scrub Jay nest predation is not af- fected by proximity to vehicle trails. In fact, jays regularly place their nests at the edges of clear- ings, including trail edges (Woolfenden 1974). Finally, seasonal activity rates of potential mammalian predators were not significantly correlated with nest predation rates (Table 2). Therefore, we conclude that although bobcats, raccoons, and other mammals undoubtedly depredate some jay nests, they appear to be less important nest predators than snakes and birds. The snakes we observed most frequently in Scrub Jay habitat at ABS were the eastern in- digo, eastern coachwhip, and southern black racer (Coluber constrictor), although virtually all the racers we saw were too small to pose a sig- nificant threat. Florida pine snakes (Pituophis melanoleucus) also were observed, but less fre- quently. All of these species are primarily di- urnal (Ernst and Barbour 1989), and 94% of the snake tracks we recorded were made during daylight. We suspect that the coachwhip, a large, locally common snake that easily traverses shrubbery and is strictly diurnal (Ernst and Bar- bour 1989), is the most frequent snake predator of the jays. As shown in Table 2, an index of diurnal predator activity is strongly correlated with nest predation rates. Because it combines data ob- tained from both visual sightings and track counts, and includes several species of birds and snakes that vary in conspicuousness, this index is at best only a crude measure of relative pred- ator activity and should be viewed with caution. Nonetheless, the correlation between the index of diurnal predator activity and nest predation is striking (Fig. 4). Activity of the snakes alone shows a significant correlation both with the progression of the jay nesting season and with jay nest-predation rates. However, activity of the diurnal bird nest predators does not cor- relate significantly with either of these features of jay nesting. Because snakes are ectotherms, their activity would be expected to increase over the course of the March-June Scrub Jay breed- ing season. Thus, the positive correlations among snake activity, date within the nesting season, and nest predation rates (Table 2, Figs. 1 and 4) suggest that, although diurnal snakes and birds are frequent nest predators, snakes are relatively more important in our system. Factors influencing nest predation.--Predators may use visual, auditory, or olfactory cues pro- vided by researchers to locate bird nests. There- fore, in studies of nesting success, it is important to determine the effect that investigator visi- tation has on the rate of nest predation (Best 1978, Gottfried and Thompson 1978, Wray et al. 1982, Westmoreland and Best 1985, Martin and Roper 1988). We addressed this problem in 1987 by visiting samples of nests on two schedules. All nests were checked in the same manner, during one season, and mostly by the same per- son. Predation rates did not differ significantly between nests visited twice daily and those vis- ited every third day. Although nest visits at three-day intervals may influence predation rates (Westmoreland and Best 1985), we con- clude that within the strictures of our experi- ment, the frequency of investigator visits had no influence on the rate of nest predation. Nest mortality usually is greater on nests with eggs than on nests with nestlings (Martin in press). However, our data show that Florida Scrub Jay nests experience significantly greater predation when they have young than when they have eggs, and this trend is consistent throughout the nesting season (Fig. 2). Adult jays visit nests more frequently when tending young than when incubating eggs (Schaub 1990). Nestling jays beg frequently and loudly, and move about in the nest. Increased activity may make nests with young easier to locate by certain predators and may contribute to the higher predation rates (Hammond and Forward 1956, Young 1963, Horn 1968, Knight and Tem- ple 1986). Furthermore, this study has impli- cated snakes as the primary nest predator. Be- cause many snakes locate prey by olfaction, in addition to vision (Ashton and Ashton 1981), increased olfactory cues emitted by nestlings might further facilitate their detection. Our results also indicate that when seasonal effects are controlled, younger nestlings expe- rience a greater rate of predation than do older nestlings (Fig. 3). This difference may reflect the decreased susceptibility of older nestlings to certain predators (e.g. Blue Jays, small snakes) and increased parental defense of nests with older nestlings (Montgomerie and Weather- head 1988). Another possibility is that the more susceptible nests are found when nestlings are young. Rates of nest predation have been found to decrease with nestling age in relatively few spe- cies (e.g. Holcomb 1972). However, several of the studies reporting nest-predation rates in- creasing with nestling age did not control for the effect of season (e.g. Young 1963, Best 1978, Woolfenden and Fitzpatrick 1984). Thus, the increased mortality on older nestlings observed in these studies may occur because older nest- lings tend to be present later in the breeding season when predation rates are higher. Nests ultimately failed because of predation more often in shrubby pastures and overgrown scrub than in open, recently burned scrub. Fitz- patrick and Woolfenden (1986) reported similar results. However, the daily rate of individual acts of predation does not differ significantly among the three habitats (Table 1). It appears that individual acts of nest predation are more likely to lead to complete nest failure in pasture and overgrown scrub than in recently burned scrub. It is unknown whether this is due to among-habitat differences in the behavior of predators, among-habitat differences in nest de- fense by jays, or both. Effects of helpers.--Florida Scrub Jay pairs as- sisted by helpers had a significantly lower rate of nest predation than pairs without helpers only during the last month of the breeding sea- son, and only for nests with nestlings (Fig. 5). Our analysis, which controlled for age and ex- perience of breeders, thus suggests that helpers were effective at reducing predation rates only when the predation pressure was at its greatest intensity (Fig. 1). How could the presence of nonbreeding helpers reduce predation on nestlings? Al- though Florida Scrub Jays are almost certainly incapable of defending their nests against at- tacks by nocturnal predators (Woolfenden and Fitzpatrick 1984), our study has shown that most nest predation is attributable to diurnal snakes and birds. Helpers could reduce the frequency of successful attacks by these diurnal predators by serving as sentinels near nests and by mob- bing potential predators once they have been detected (Francis et al. 1989, McGowan and Woolfenden 1989, Mumme in press). The finding that helpers did not significantly reduce the rate of predation on eggs is not sur- prising; nonbreeders normally do not partici- pate in reproductive activities until after hatch- ing. In fact, breeders usually chase potential helpers away from the immediate vicinity of the nest during laying and incubation (Stallcup and Woolfenden 1978). Furthermore, non- breeders have been seen to remove eggs from the nests of the pairs with which they are as- sociated (Woolfenden 1974). Although the re- sults of Woolfenden and Fitzpatrick (1984) sug- gest that the presence of helpers reduces predation on eggs, our analysis--which em- ployed a more extensive data set and controlled for the potentially confounding effects of sea- son, breeder experience, and senescence--shows no such effect. Woolfenden and Fitzpatrick (1984:204) re- ported that helpers significantly enhance sur- vival of eggs and nestlings only during the sea- son's first nesting attempt. For seasonal renests, helpers had no significant influence on the sur- vival rates. Our analysis differs from theirs in at least one important respect; we investigated seasonal influences by examining nest preda- tion in monthly and half-monthly intervals, re- gardless of whether a particular nest was a first nest or a tenest. Our analysis indicates that helpers reduced nest predation only during the final month of the nesting season. We propose that the results of Woolfenden and Fitzpatrick (1984), suggesting that helpers enhance repro- duction of the first attempts, may have been confounded by correlated differences in breed- er experience. Pairs without helpers are fre- quently inexperienced breeders that do not be- gin nesting until late in the season, when nest predation rates are high and success rates are low (Woolfenden and Fitzpatrick 1984:216). Thus, the higher survival observed in seasonal first nests produced by pairs with helpers may be attributed to breeder experience rather than to the effects of helpers per se. In contrast, the analysis reported in this paper examined helper effects while controlling for age and experience of breeders. ACKNOWLEDGMENTS We are grateful to the staff of Archbold Biological Station for providing access to the Station's outstand- ing research facilities. We also thank Jack P. Hailman, Thomas E. Martin, Earl D. McCoy, and Henry R. Mu- shinsky for their constructive criticisms of earlier ver- sions of the manuscript. Financial support was pro- vided by NSF grant BSR 86-00174 to Mumme, and an ABS Grant-in-Aid and Sigma Xi grant to Schaub. LrrERtURE CITED ASHTON, R. E., AND P.S. ASHTON. 1981. Handbook of reptiles and amphibians of Florida; the snakes. Windward Publishing, Inc., Miami, Florida. BEST, L.B. 1978. Field Sparrow reproductive success and nesting ecology. Auk 95:9-22. ERNST, C. H., AND R. W. BARBOUR. 1989. Snakes of eastern North America. George Mason Univ. Press, Fairfax, Virginia. FITZPATRICK, J. W., AND G. E. WOOLFENDEN. 1986. De- mographic routes to cooperative breeding in some New World jays. Pages 137-160 in Evolution of animal behavior (M. H. Nitecki and J. A. Kitcheil, Eds.). Oxford Univ. Press, New York. FRANCIS, A. M., J. P. HAILMAN, AND G. E. WOOLFENDEN. 1989. Mobbing by Florida Scrub Jays: Behaviour, sexual asymmetry, role of helpers and ontogeny. Anita. Behar. 38:795-816. GOTTFRIED, B. M., AND C. F. THOMPSON. 1978. Ex- perimental analysis of nest predation in an old- field habitat. Auk 95:304-312. HAMMOND, M. C., AND W. R. FORWARD. 1956. Ex- periment on causes of duck nest predation. J. Wildl. Manage. 20:243-247. HOLCOMB, L.C. 1972. Nest success and age-specific mortality in Traill's Flycatcher. Auk 89:837-841. HORN, H. S. 1968. The adaptive significance of co- lonial nesting in the Brewer's Blackbird Euphagus cyanocephalus. Ecology 49:682-694. KEPPIE, D. g., P. W. HERZOG. 1978. Nest site characteristics and nest success of Spruce Grouse. J. Wildl. Manage. 42:628-632. KNIGHT, R. L., AND S. A. TEMPLE. 1986. Nest defense in the American Goldfinch. Anita. Behav. 34:887- 897. LOISELLE, g. A., AND W. G. HOPEES. 1983. Nest pre- dation in insular and mainland lowland rainfor- est in Panama. Condor 85:93-95. MARTIN, T.E. 1987. Artificial nest experiments: Ef- fects of nest appearance and type of predator. Condor 89:925-928. MARTIN, T. E. 1988a. Nest placement: Implications for selected life-history traits, with special ref- erence to clutch size. Am. Nat. 132:900-910. MARTIN, T. E. 1988b. Processes organizing open- nesting bird assemblages: Competition or nest predation? Evol. Ecol. 2:37-50. MARTIN, T.E. In press. Breeding productivity con- siderations: What are the appropriate habitat fea- tures for management? In Ecology and conser- vation of Neotropical migrant birds (J. M. Hagan and D. W. Johnston, Eds.). Smithsonian Institu- tion Press, Washington, D.C. MARTIN, T. E., J. J. ROPER. 1988. Nest predation and nest-site selection of a western population of the Hermit Thrush. Condor 90:51-57. McGowAN, K. J., ND G. E. WOOLFENDEN. 1989. A sentinel system in the Florida Scrub Jay. Anita. Behav. 37:1000-1006. MONTGOMERm, R. D., ND P. J. WEATHERHEAD. 1988. Risks and rewards of nest defense by parent birds. Q. Rev. Biol. 63:167-187. MUMME, R. L. 1987. Eastern indigo snake preys on juvenile Florida Scrub Jay. Fla. Field Nat. 15:53- 54. MUMME, R. L. In press. Helping behaviour in the Florida Scrub Jay: Nonaptation, exaptation, or ad- aptation? Pages 1317-1324 in Acta XX Congressus Internationalis Ornithologici. Christchurch, New Zealand, 1990. New Zealand Ornithol. Congr. Trust Board, Wellington. SCHAU, R. 1990. Predation on the eggs and nest- lings of Florida Scrub Jays. M.S. thesis, Univ. South Florida, Tampa. SLAGSVOLD, T. 1982. Clutch size variation in passer- ine birds: The nest predation hypothesis. Oecolo- gia 54:159-169. SOKAL, R. R., NI F. J. ROHLF. 1981. Biometry, 2nd ed. W. H. Freeman, New York. STALLCUP, J. A., AND G. E. WOOLFFaqDEN. 1978. Family status and contributions to breeding by Florida Scrub Jays. Anita. Behav. 26:1144-1156. WASSMER, D. A., D. D. GUENTHER, AND J. N. LAYNE. 1988. Ecology of the bobcat in south-central Flor- ida. Bull. Fla. State Mus., Biol. Sci. 33(4):159-228. WEER, T.A. 1980. Easterncoachwhippredationon juvenile Scrub Jays. Fla. Field Nat. 8:29-30. WESTCOTT, P. W. 1970. Ecology and behavior of the Florida Scrub Jay. Ph.D. dissertation, Univ. Flor- ida, Gainesville. WESTMORELAND, D., AND L. B. BEST. 1985. The effect of disturbance on Mourning Dove nesting suc- cess. Auk 102:774-780. WILCOVE, D.S. 1985. Nest predation in forest tracts and the decline of migratory songbirds. Ecology 66:1211-1214. WOOLFENDEN, G.E. 1974. Nesting and survival in a population of Florida Scrub Jays. Living Bird 12: 25-49. WOOLFFaqDEN, G. E. 1978. Growth and survival of Florida Scrub Jays. Wilson Bull. 90:1-18. WOOLFFaqDEN, G. E.,ANDJ. W. FITZPATRICK. 1984. The Florida Scrub Jay: Demography of a cooperative- breeding bird. Princeton Univ. Press, Princeton, New Jersey. WOOLFENDEN, G. E., AND J. W. FITZPATRICK. 1990. Florida Scrub jays: A synopsis after 18 years of study. Pages 239-266 in Cooperative breeding in birds (P. B. Stacey and W. D. Koenig, Eds.). Cam- bridge Univ. Press, Cambridge. WOOLFEN'DEN, G. E., _NI J. W. FITZPATRICK. 1991. Florida Scrub Jay ecology and conservation. Pages 542-565 in Bird population studies: Relevance to conservation and management (C. M. Perrins, J.- D. Lebreton, and G. J. M. Hirons, Eds.). Oxford Univ. Press, Oxford. WORtEY, D. J. 1980. Nest sites, movement, and ac- tivity patterns of the raccoon Procyon lotor in south- central Florida. M.A. thesis, Univ. South Florida, Tampa. WRAY, T., II., K. A. STRrr, AND R. C. WHITMORE. 1982. Reproductive success of grassland sparrows on a reclaimed surface mine in West Virginia. Auk 99: 157-164. YOUNG, H. 1963. Age-specific mortality in the eggs and nestlings of blackbirds. Auk 80:145-155. ZAR, J.H. 1984. Biostatisticalanalysis. Prentice-Hall, Englewood Cliffs, New Jersey.