USING VITELLOGENIN TO IDENTIFY INTERANNUAL VARIATION IN BREEDING CHRONOLOGY OF MARBLED MURRELETS (BRACHYRAMPHUS MARMORATUS)

USING VITELLOGENIN TO IDENTIFY INTERANNUAL VARIATION IN BREEDING CHRONOLOGY OF MARBLED MURRELETS (BRACHYRAMPHUS MARMORATUS) Author(s): Laura Mcfarlane Tranquilla, Tony Williams, Fred Cooke Source: The Auk, 120(2):512-521. Published By: The American Ornithologists' Union DOI: http://dx.doi.org/10.1642/0004-8038(2003)120[0512:uvtiiv]2.0.co;2 URL: http://www.bioone.org/doi/full/10.1642/0004-8038%282003%29120%5b0512%3auvtiiv %5D2.0.CO%3B2 BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder. BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.

The Auk 120(2):512 521, 2003 USING VITELLOGENIN TO IDENTIFY INTERANNUAL VARIATION IN BREEDING CHRONOLOGY OF MARBLED MURRELETS (BRACHYRAMPHUS MARMORATUS) LAURA MCFARLANE TRANQUILLA 1, TONY WILLIAMS, AND FRED COOKE 2 Centre for Wildlife Ecology, Department of Biological Sciences, Simon Fraser University, 8888 University Drive, Burnaby, British Columbia V5A 1S6, Canada ABSTRACT. Vitellogenin is a lipophosphoprotein found in plasma of egg-producing birds prior to laying that may be used to identify fecund females whose reproductive status is otherwise unknown. We captured Marbled Murrelets (Brachyramphus marmoratus) at sea in Desolation Sound, British Columbia, and used vitellogenin to (1) identify variation in egg production between 1999 and 2000, (2) predict timing of subsequent breeding stages on the basis of egg production, and (3) describe proportion of captured females producing eggs. We also used vitellogenin to investigate a capture bias previously detected in mist-netted birds in the study area and found a corresponding bias in number of egg producers caught. Dates that egg producers were present (27 April to 6 July 1999, 20 April to 6 July 2000) indicate that breeding is highly asynchronous in that species but was similar in both years. Predicted chick-fledging based on vitellogenin analyses was within one day of first sightings of fledglings at sea in both years, confirming that the vitellogenin technique provides accurate information on breeding chronology. Percentage of egg producers (54% in 1999, 56% in 2000) were similar in both years. Vitellogenin analyses provided a chronology very similar to that previously estimated using multiple techniques in the same study area (1996 1998), confirming that vitellogenin analyses alone may be used to describe chronology when sampling encompasses the entire laying period. We recommend that technique for use in other studies of secretive species where egg production cannot normally be monitored by direct observation. Received 3 January 2002, accepted 14 December 2002 RÉSUMÉ. La vitellogénine est une lipophosphoprotéine que l on retrouve dans le plasma des œufs produits par les oiseaux avant la ponte. Cette protéine peut être utilisée pour identifier les femelles fécondes dont le statut reproducteur resterait inconnu autrement. Nous avons capturé des Guillemots marbrés (Brachyramphus marmoratus) en mer à Desolation Sound, Colombie- Britannique, et nous les avons utilisé pour (1) identifier les variations dans la production d œufs entre 1999 et 2000, (2) prédire le déroulement des étapes de reproduction subséquentes en se basant sur la production d œufs. Nous avons également utilisé la vitellogénine pour examiner le biais de capture précédemment détecté chez les oiseaux capturés au filet dans l aire d étude. Nous avons ainsi décelé ce biais dans le nombre d oiseaux capturés produisant des œufs. Les dates durant lesquelles ces oiseaux étaient présents (du 27 avril au 6 juillet 1999, du 20 avril au 6 juillet 2000) indiquent que la reproduction est fortement asynchrone chez cette espèce mais elle a été similaire pour les deux années. En se basant sur les analyses de vitellogénine, la prédiction du nombre d œufs pondus correspondait aux nombres de jeunes recensés au cours des premières observations en mer et ce pour les deux années. Ceci vient confirmer que la technique de la vitellogénine fournit une information précise sur la chronologie de reproduction. Le pourcentage d oiseaux produisant des œufs (54% en 1999, 56% en 2000) étaient similaires dans les deux années. Les analyses de vitellogénine ont fourni une chronologie très similaire à celle précédemment estimée à partir de l utilisation de techniques multiples pour la même aire d étude (1996 1998), confirmant que les analyses de vitellogénine peuvent être utilisées comme technique unique pour décrire la chronologie quand l échantillonnage porte sur la totalité de la période de ponte. Nous recommandons l utilisation de cette technique dans les autres études traitant d espèces dont la production d œufs ne peut pas être suivie normalement par observations directes. 1 E-mail: lat@sfu.ca 2 Present address: Larkin s Cottage, 6 Lynn Road, Castle Rising, Norfolk PE31-6AB, United Kingdom. 512

April 2003] Breeding Chronology in Marbled Murrelets 513 STUDYING THE MARBLED Murrelet (Brachyramphus marmoratus) has been challenging because of its secretive habits (Cooke 1999). Reproductive status of individuals is especially difficult to determine, because that species nests solitarily and cryptically on mossy limbs of old-growth trees (Nelson 1997), making it difficult to determine precisely timing of breeding activities such as laying or incubation. Nevertheless, accurately evaluating activity patterns and habitat use requires an understanding of breeding chronology (Hamer and Nelson 1995). Breeding chronologies were first estimated using Marbled Murrelets collected at sea (Sealy 1974, 1975; Carter 1984). That work examined variation in reproductive tract morphology with respect to brood patch development, fish-holding by adults, and appearance of fledged young at sea, to estimate breeding phenology for Marbled Murrelets in British Columbia. Although successful, dissection of breeding birds was required and with classification of the Marbled Murrelet as a threatened or endangered species, collecting is no longer an option. More recently, direct observations at the nest have been the only means of accurately determining breeding status of individuals; however, that approach has not been able to assess overall reproductive status of the local population, because it depends on observations of individuals already known to be breeding and has been limited by small sample sizes. Due to their cryptic nesting habits, Marbled Murrelets are more easily studied when off the nest, at sea (Vanderkist et al. 2000, Speckman et al. 2000). However, those sexually monomorphic birds cannot be aged past first year and age of first breeding is unknown (Nelson 1997). Thus, characteristics of plumage and measurements of body size offer no clues to identifying breeding status and dissection of reproductive tracts is no longer an option (Sealy 1974). As well, brood patch scores have not always reliably matched breeding status (McFarlane Tranquilla 2001). Nondestructive analyses of plasma sampled from birds when captured may, however, be used to identify fecund females, using an egg yolk protein called vitellogenin (Vanderkist et al. 2000). Vitellogenin (VTG) is a lipophosphoprotein secreted by the liver in response to estrogenic stimulation and is transported to the ovary via the bloodstream and deposited in the developing oocyte (Deeley et al. 1975, Wang and Williams 1982). Elevated levels of VTG in the plasma coincide with yolk deposition (Challenger et al. 2001) and are undetectable in physiologically immature birds, mature females during the nonbreeding season, and males (Deeley et al. 1975, Mitchell and Carlisle 1991). Thus, measurements of VTG levels in plasma accurately identify fecund females (Mitchell and Carlisle 1991). Given the potentially widespread use of that technique, the first aim of this article was to further assess the utility of VTG analysis, including effects of freezing and of assaying plasma from birds whose reproductive status was not known. Second, the two main techniques for capturing Marbled Murrelets dip netting and mist netting catch different groups of birds; mist netting captures a male-biased sample, whereas dip netting does not (Vanderkist et al. 1999). We therefore used VTG analyses to investigate that capture bias, with the expectation that egg-producing females would not be captured in mist nets (Vanderkist et al. 1999, Bradley et al. 2002). Finally, we compared our estimates of breeding chronology made using VTG data with Lougheed et al. s (2002), who used several techniques (including VTG analyses) to compare midpoints of breeding stages, construct a breeding chronology, and assess interannual variation in chronology for Marbled Murrelets in Desolation Sound. We extended Vanderkist et al. s (1999) study, which collected only late-season VTG data, to (1) identify interannual variation in breeding chronology using VTG analysis to delimit duration of egg production, (2) predict incubation period and fledging dates by extrapolation (following Lougheed et al. 2002), and (3) estimate proportion of egg-producing females. METHODS Study site and blood sampling. Marbled Murrelets were captured in Desolation Sound, British Columbia (50 05 N, 124 40 W), from 20 April to 4 September 1999 and from 19 April to 26 August 2000. Birds were dip netted (Whitworth et al. 1997, Vanderkist et al. 1999) at night (2100 0400 hours PST) in Desolation Sound and mist netted (8 June to 30 July 1999, 14 June to 29 July 2000) at the mouth of Theodosia Inlet, Desolation Sound (50 04 N, 124º42 W) (Kaiser et al. 1995). Prior to 1 June, we could not mist-net murrelets because they did not use that inlet as a flyway to their nest (Kaiser et al. 1995), but after 1 June, birds were captured using both mist-net and dip-net techniques. Dip netting occurred on 108 of 137 days in 1999 and 96 of 129 days

514 TRANQUILLA, WILLIAMS, AND COOKE [Auk, Vol. 120 in 2000, and mist-netting occurred on 41 of 52 days in 1999 and 37 of 45 days in 2000. Days were missed due to inclement weather. Marbled Murrelets captured by both methods were bled, beginning early and extending well past the known egg-laying period (early May to mid July, Lougheed et al. 2002). Birds were bled with no a priori knowledge of their sex and samples were later sexed using genomic DNA (Griffiths et al. 1996, Vanderkist et al. 1999). Blood samples (1 2 ml) were taken from the brachial vein, dispensed into an Eppendorf tube, and kept cold until they were returned to field camp. Within 6 8 h of collection, blood was centrifuged at 6,000 rpm for 10 min, plasma was removed, and red blood cells and plasma were frozen at 20 C until they could be transported to the laboratory for further analysis. Vitellogenin analyses. Marbled Murrelets lay only one egg per breeding attempt (Sealy 1974); therefore, plasma VTG in egg producers was expected to decline immediately after the yolk of that egg was completed (e.g. Challenger et al. 2001). Vitellogenin concentration in the plasma was determined indirectly, with an assay for vitellogenic zinc (following Mitchell and Carlisle 1991, Vanderkist et al. 2000). Vitellogenic zinc (VTG-Zn) was used as an index for VTG, as described and validated for Marbled Murrelets by Vanderkist et al. (2000). We investigated two possible sources of error in assessing fecundity using VTG analysis: (1) effect of dilution ratio for birds of unknown status and (2) effect of long-term storage at 20 C. Following Mitchell and Carlisle (1991), the VTG assay should be adjusted according to whether the plasma is from nonlaying or laying birds, by diluting plasma from nonlaying birds 2-fold and from laying birds 4-fold. However, due to the murrelets unknown status, we analysed all plasma as if it was from laying birds (i.e. diluted 4-fold). That would not adversely affect analyses of plasma that actually was from egg-producing Marbled Murrelets, as the assay would have been done correctly. For males and nonlaying females, doing the assay in that way (i.e. at the wrong dilution) produced negative values for approximately half of the capture sample, which were not theoretically possible. Thus, after sex and laying status were determined, we corrected negative values by reassaying 91 subsamples of plasma from males and non-eggproducers, at the correct 2-fold dilution. Similarly, to assess effect of long-term storage, we assayed 14 subsamples of plasma after 5 months and then again after 18 months of storage at 20 C. DNA sexing. Marbled Murrelets were sexed following the methods described in Vanderkist et al. (1999), with the modification that blood samples were either centrifuged red blood cells or blood cells dried on filter paper. Some birds could not be sexed and were omitted from analyses that required sex as a variable. Classifying egg producers. We used the term egg producers because it was unclear how many days elapsed between producing and laying an egg, or whether producers (as defined by elevated VTG) always became layers. As well, VTG data only identified producers, and not whether eggs were from first or second clutches. Using VTG data cannot lead to any assumptions about the success of the breeding attempt following egg laying. To establish a VTG threshold that would identify egg-producing females, we measured the amount of VTG in birds that clearly were not producing eggs (i.e. males). Following Vanderkist et al. (1999), we constructed limits around mean VTG + 3 SD units for 103 males (1999 and 2000 pooled data) (µ SE = 0.26 0.03; + 3 SD = 0.96 µg ml 1 VTG). We did the same for 41 females that probably were not producing eggs, because they were caught at the end of the season, in August; that VTG value was not significantly different from that of males (µ SE = 0.30 0.03; + 3 SD = 0.90 µg ml 1 VTG). Thus, we concluded that VTG values >0.96 µg ml 1 were outside 99.4% (µ + 3 ) of a normal distribution (see Zar 1996) and should have been from birds that were producing eggs. VTG was measured in micrograms per milliliter (µg ml 1 ). Constructing breeding chronology. We used plasma VTG data to construct a breeding chronology of Marbled Murrelets in Desolation Sound. Because Marbled Murrelets lay asynchronously, with incubation and chick stages spread over ~4 months, we calculated core periods when most birds were at a particular breeding stage (see Lougheed et al. 2002). Thus, mean dates plus one SD from the mean (i.e. 68% of the data; Zar 1996) were used to describe core breeding stages. Mean date of egg production was calculated using only egg-producing females (Fig. 2, above hatched line). Mean egg-production periods were used to predict subsequent breeding stages, including the first appearances of juveniles. Following Lougheed et al. (2002), a 30 day incubation period (Nelson 1997) and a 28 day chick-rearing period (Hamer and Nelson 1995) were used to calculate breeding chronology. In the closely related Cassin s Auklet (Ptychoramphus aleuticus), eggs require 14 days for production (Astheimer 1986). Thus, because elevated VTG coincides with rapid yolk development early in egg production (Challenger et al. 2001), we estimated that birds with elevated VTG were halfway through egg production. Therefore, individual laying dates were estimated to occur seven days following capture of birds with elevated VTG (Lougheed et al. 2002). We assumed there was no capture effect that might extend the number of days between egg production and laying beyond 14 days. Nighttime detections of juvenile Marbled Murrelets (Centre for Wildlife Ecology unpubl. data) were compared with the fledging chronology predicted from VTG data. From April to August, dip netting

April 2003] Breeding Chronology in Marbled Murrelets 515 occurred on 84% of the nights, whereas at-sea surveys were conducted on 6% of the days (every 10 days when weather permitted). Thus, the dip-net effort was much more constant than the corresponding daytime surveys. Although the detection radius for juvenile murrelets was probably smaller at night than during the day, overall coverage of the study area was greater and more constant using the dip-net method. As well, using the juveniles seen while dip netting maintained consistent study methodology between those and the egg-producing females. Proportions of egg producers. Proportions of egg producers (out of all females) in both years were calculated for each 20 day SD from the mean eggproducing date. Proportions of egg producers were calculated from 17 April (2 SD before the mean) to 15 August (5 SD after the mean). Data analysis. Statistical analyses were conducted using MINITAB (version 13; Minitab 2000) software. Differences in plasma VTG concentrations (1) at 4- fold versus 2-fold dilutions and (2) before versus after long-term freezing were tested using paired t-tests. All plasma was not available for reassaying to test the dilution or freezing effects; thus, after determining that differences in values of (1) and (2) were significant, ordinary least squares (OLS) regressions (Ricker 1984) of those values were used to predict VTG values for the remaining plasma samples that could not be reassayed. Normality was assessed using Anderson-Darling normality tests and nonparametric tests were used for differences in VTG between mist-netted and dip-netted birds (Mann-Whitney U-test) and median dates of egg production (Kruskall-Wallis test). As mean egg-producing dates did not differ significantly between years, pooled mean egg-producing date was used when comparing proportions of egg producers. Proportions of breeders in 20 day intervals around mean egg-producing date were assessed using a test and confidence interval for two proportions. Mean period for each breeding stage indicates the mean date SD. Significance was assumed at = 0.05. VTG values are given in mean SE except where noted otherwise. RESULTS Methodological considerations. Assaying plasma from males at any breeding stage and from chick-rearing females at the wrong dilution (i.e. laying female) generated negative VTG values in approximately half of the captured birds. Reassaying those samples at the correct (nonbreeding) dilution corrected that problem and increased mean VTG in nonbreeders from 0.14 0.40 SD to 0.23 0.20 SD. As we could not re-assay all of the nonbreeding samples, an OLS regression of paired values (of 4- vs. 2-fold dilutions, n = 91) for total zinc and depleted zinc (sensu Mitchell and Carlisle 1991) was used to correct the remaining data from non-egg-producing birds prior to subsequent analysis (OLS regression equations for 4- vs. 2-fold dilutions: total plasma Zn = 0.4837 + 0.6297(x), depleted plasma Zn = 0.6909 + 0.3761(x)). Long-term freezing caused a significant decrease in VTG (Table 1). As with the negative values above, a regression of those paired values was used to calculate the theoretical effect of long-term freezing for the entire sample of egg producers (n = 63), mimicking an 18 month freezing time. However, although the time-dependent decrease in plasma VTG was significant, all 63 females still would have been classified as egg producers if those samples had been assayed after 18 months of storage (based on VTG > 0.96 µg ml 1 ). That is, the decrease in VTG was not enough to reclassify egg-producing plasma as non-egg-producing plasma. Comparison of capture methods. In both years, mean VTG was significantly higher in dipnetted females than in mist-netted females (Fig. 1). In 1999, when mist-netting and dip netting occurred at the same time, 50% (10/20) of females dip netted were producing eggs, whereas none of the five females mist netted was producing an egg. In 2000, 45% (13/29) of females dip netted were producing eggs, whereas none of the six females mist-netted was producing an egg. Distribution of egg producers. The duration of the egg-producing phase in the population was delimited by plotting VTG values in all females captured in 1999 and 2000 separately (Fig. 2). Between 20 April and 4 September 1999, blood samples from 92 females were analysed for VTG (mean = 1.37 0.24 µg ml 1 ). Of those, 34% (31/92) had elevated VTG ( 0.96; mean = TABLE 1. Decrease in detectable VTG in plasma with time and freezing in egg-producing Marbled Murrelets. Data are ± SE. Approximate Average Average Average final time in freezer total Zn depleted Zn VTG-Zn 5 months (n = 14) 7.03 ± 0.40 1.77 ± 0.15 5.26 ± 0.35 18 months (n = 14) 4.63 ± 0.39 1.79 ± 0.13 2.85 ± 0.47 Paired t-test P = 0.001 P = 0.84 P = 0.001

516 TRANQUILLA, WILLIAMS, AND COOKE [Auk, Vol. 120 4.087 0.36 µg ml 1 ). Egg producers were captured from 27 April to 6 July 1999 (Fig. 2). Mean egg-producing date in 1999 was 24 May 20 days (SD), modal date was 9 May (Fig. 3A), and estimated duration of egg production was 70 days. Between 19 April and 31 July 2000, blood samples from 66 females were analysed for VTG (mean = 2.05 0.35 µg ml 1 ). Of those, 49% (32/66) had elevated VTG ( 0.96; mean = 4.20 0.44 µg ml 1 ). Egg producers were caught from 20 April through 6 July 2000 (Fig. 2). Mean eggproducing date in 2000 was 29 May 21 days (SD), modal date was 24 May (Fig. 3B), and estimated duration of egg production was 77 days. Median dates of egg production did not differ between years (Kruskal-Wallis H = 0.82, df = 1, P = 0.37). Mean date of egg production for pooled years was 27 May 2.5 days (SE). Predicting breeding chronology. We used mean dates of egg production to construct breeding chronology by predicting successive breeding stages. We found breeding chronology to be similar between years. Mean eggproducing period (years pooled) ranged from 7 May to 16 June (Table 2). First egg producers were captured earlier than by Vanderkist et al. (1999); however, mean incubation and chick rearing periods were similar to Lougheed et al. s (2002). Based on the first egg producers (see Table 2), we predicted the first juveniles would appear at sea on 1 July 1999 and 24 June 2000. First juveniles were actually seen at night on 30 June and 25 June, respectively (Table 2). FIG. 1. Plasma levels of yolk precursor (VTG-Zn (µg ml 1 )) found in Marbled Murrelets, by sex (dark = F, light = M) and capture method (hatched = mist net, open = dip net), years pooled. Data include birds captured during the entire breeding season (April to August). Asterisk (*) indicates group that is significantly different (Mann-Whitney U-test, P = 0.001). FIG. 2. Vitellogenin levels in females by date, describing duration of egg-producing period for two breeding seasons (birds captured from April to August September) in Marbled Murrelets in Desolation Sound, British Columbia. Values above 0.955 (dotted line) indicate egg-producing birds. Mean egg-producing date is shown as µ = (x) SD and is calculated using egg-producing birds (i.e. above dotted line). Lougheed et al. (2002) recorded a very early juvenile in 1998; apart from that early sighting, the first recorded juveniles in our study fall on the dates expected from previous years of at-sea surveys (see Table 2). The last juvenile was predicted to fledge on 9 September in both years. Calculating from first (20 April) and last (6 July) egg producers resulted in a fledging span of 77 days, from 24 June to 9 September. Extrapolating from the dates for first and last egg producers the entire breeding season spanned 142 days. Proportions of egg producers. No egg producers were captured after 6 July in either year. Proportions of egg producers in each 20 day SD from the mean did not differ significantly between the two years (Table 3). Proportions were highest between 8 May and 27 May, ranging from 0.42 to 0.87 egg-producing females, but did not differ significantly until they dropped after 7 July to 0.008 0 (P = 0.001). Averaging the entire egg production period (April to July), proportions of egg-producing females captured were 0.54 (n = 58, 1999) and 0.56 (n = 57, 2000). DISCUSSION Comparison of capture methods. Methods of capture usually assume random sampling of the population. However, that assumption has

April 2003] Breeding Chronology in Marbled Murrelets 517 TABLE 2. Breeding chronology with predictions for timing of breeding stages in Marbled Murrelets. Data are ± SD (days). Included are data from Vanderkist et al. (1999) for comparison to a previous study using VTG and Lougheed et al. (2002) for comparison of estimating chronology using multiple methods. All studies were conducted in Desolation Sound, British Columbia. First egg Last egg Mean egg Mean predicted Mean Mean First First Start of producer producer producing incubation predicted chick predicted predicted juvenile Study Year captures caught caught period period rearing period fledging period a juvenileb seenc This study 1999 20 April 27 April 6 July 24 May 19 June 1 19 July 1 19 28 July 20 1 July 30 June (D) 2000 19 April 20 April 6 July 29 May 21 June 5 21 July 5 21 2 August 20 24 June 25 June (D) Pooled 27 May 20 June 3 20 July 3 20 31 July 20 Lougheed et al. d 1996 19 May 8 July 18 June 4 Aug 25 June (S) 1997 27 June (S) 1998 11 June (S) Vanderkist et al. 1997 14 May 21 May 3 July 25 July 27 June (S) a Mean predicted fledging period = (mean egg laying period + SD) + 7 days to egg lay + 30 days (incubation) + 28 days (chick rearing). b First juvenile predicted = date first egg producer caught + 7 days to egg lay + 30 days (incubation) + 28 days (chick rearing). c D indicates while dipnetting (night), and S indicates on at-sea surveys (day). d Dates describe core periods, that contain the middle 50% of the data and averaged over three years (from tables 1 and 2 in Lougheed et al. 2002). FIG. 3. Distribution of egg producers by date in (A) 1999 (n = 31) and (B) 2000 (n = 32) (bars are grouped in 10 day periods). been challenged due to a significant male bias found in Marbled Murrelets caught in mist nets in Desolation Sound (Vanderkist et al. 1999, Bradley et al. 2002). It has been suggested that mist-netting (unlike dip netting) catches Marbled Murrelets only during the chick-rearing stage (Vanderkist et al. 1999) and that the bias in favor of males results from increased inland flights by males carrying fish to their young (Bradley et al. 2002). Our results support that hypothesis, as no egg-producing murrelets were caught in the mist nets. Vitellogenin in dip-netted females was significantly higher overall, supporting the hypothesis that mistnetted females were from a different subset of the population, or used the area only after laying. Those results support Vanderkist et al. s (1999) suggestion that different capture methods cannot be assumed to sample populations randomly. The egg-producing period in the Marbled Murrelet. Duration of egg laying in Marbled Murrelets has been estimated from early May to mid-july in British Columbia (Sealy 1975, Carter 1984). The present study extends the earliest laying estimate by about two weeks and coincides closely with the range of laying dates found by Lougheed et al. (2002) in Desolation Sound in 1996 1998. On the Queen Charlotte Islands,

518 TRANQUILLA, WILLIAMS, AND COOKE [Auk, Vol. 120 TABLE 3. Percentages of egg producing female Marbled Murrelets dipnetted in 1999 and 2000. Percentages in each period did not differ by year (i.e. across columns). Percentage of females that were producing eggs Dates a 1999 2000 17 April to 7 May 41.7% (12) b 83.3% (6) A c 8 May to 27 May 67.7% (21) 87.5% (16) A 28 May to 16 June 41.7% (12) 47.1% (17) A 17 June to 6 July 54.6% (11) 50% (16) A 7 July to 26 July 0% (6) 0% (10) B 27 July to 15 August 0% (19) 0% (2) B a Each period is 1 SD (i.e. each lasting 20 days) around the mean egg laying date, in bold (see Figure 2). b Sample sizes are in parentheses. c Rows with the letter A are not significantly different from each other. Rows with the letter B are significantly different from those with the letter A (P = 0.00 in 1999, P = 0.01 in 2000). north of Desolation Sound, Sealy (1975) estimated the laying period to be 42 49 days, from mid-may to late June or early July. Lougheed et al. (2002) calculated the egg-production period to be 79 days; our results were almost identical, with an estimate of 77 days. The dates juveniles appeared at sea, predicted from dates of egg production, closely matched the actual dates of juvenile appearance in 1999 and 2000. In 1999, birds were captured for 7 days before any egg producers were identified (Table 2), which suggests that onset of egg production was not missed that year. However, in 2000, capture of egg-producing females only one day after captures began suggests that VTG sampling may have missed the earliest egg producers. If true, our predicted first juvenile appearance in 2000 is later than it should be; however, the close timing of our prediction with actual juvenile appearance in 2000 refutes that possibility. If early egg producers were missed, both the comparison with Lougheed et al. s (2002) breeding chronology, and the insignificant differences between the frequency distributions and the proportions of egg producers in our two years of study, suggest that that has resulted in a relatively minor discrepancy in our chronology. Marbled Murrelets at our site, as in other populations (Nelson 1997), produced eggs asynchronously, from April to July. That has generally been attributed to their solitary nesting habits (Hamer and Nelson 1995, Gaston and Jones 1998). However, it still is unclear whether nesting asynchrony is promoted by renesting after clutch failure or to variation in clutch initiation among individuals (Nelson 1997). Frequency of renesting in Marbled Murrelets is not known, but in light of the long breeding season, renesting probably plays an important role in breeding asynchrony in that species. As well, renesting would be an important adaptation for Marbled Murrelets, whose nests appear highly vulnerable to predation (Nelson and Hamer 1995). Other alcids lay eggs more or less synchronously in different years depending on species and location (Ainley and Boekelheide 1990, Gaston and Jones 1998). In alcids, trends generally reflect a latitudinal cline in asynchrony, with breeding seasons shortening with increasing latitude (Lougheed et al. 2002). Marbled Murrelets, however, are more asynchronous than would be predicted from their breeding latitude (Lougheed et al. 2002). Interannual variation in breeding chronology may thus be more difficult to detect in Marbled Murrelets than in birds with a shorter breeding season, when laying occurs over a shorter period of time. Also, factors affecting individual variation in onset of breeding (e.g. Sydeman et al. 1991, Wiemerskirch 1992, Gaston et al. 1994) are poorly understood for Marbled Murrelets. Selective forces that shape breeding chronology in so many other seabirds are not yet clear for Marbled Murrelets. Breeding chronology. Although females captured well before or after they were producing eggs would not have been detected by VTG analyses, the duration and range of egg production should be accurate, assuming an equal probability of catching any female throughout the season. Mean date and duration of egg production estimated using the VTG method were nearly identical in both years of our study. By contrast, Lougheed et al. (2002) recorded earlier laying at the same site from 1996 through 1998 which coincided with significant increases of sea-surface temperature from year to year during the breeding season. The trend detected by Lougheed et al. (2002) may not have continued into the years of our study. Alternatively, differences between the results in Lougheed et al. (2002) and our study could be attributable, in part, to differences in methodology. Lougheed et al. (2002) used four different methods to construct the breeding chronology and, in doing so, some calculations produced different results for the same year (e.g. different estimates for onset of incubation).

April 2003] Breeding Chronology in Marbled Murrelets 519 Our single method may have more consistently assessed interannual variation. Our captures, which ranged over the same dates for both years of the study, may have influenced the early end of the data distribution in 2000; however, that was not due to flaws in VTG analysis, but to our sampling methodology. In future, researchers should begin capturing Marbled Murrelets in March and early April to ensure that the earliest egg producers are detected. The duration of our predicted breeding season (20 April to 9 September) compared closely to that described by Lougheed et al. (2002) (21 April to 5 September). Lougheed et al. s (2002) estimates of chronology derived from counts of juveniles at sea were significantly earlier than those derived from VTG analyses; however, the VTG data she used (from Vanderkist 1999) were collected later in the season, which missed the early egg producers. As well, Lougheed et al. (2002) noted that counts of juveniles at sea were truncated near the end of the season, because later-fledged juveniles could not be distinguished from moulting adults. Juvenile counts and VTG analyses are complementary and may be used to describe breeding chronology more completely and precisely. If the earliest eggproducing females are not sampled, juvenile counts at sea would provide an estimate of the earliest egg-production; whereas the last VTG sampling would estimate the last individuals that fledge but that are overlooked during atsea surveys late in the season. Because Marbled Murrelets lay asynchronously, we presented mean periods as conservative estimates during which researchers could be more confident of detecting birds (e.g. during an at-sea survey) in particular breeding periods. The diversity of methods described in Lougheed et al. (2002) provided an accurate description of the eggproduction period and could be used in other studies to describe the timing of egg production when VTG analyses are not possible. Likewise, one method can accurately predict all breeding stages if it produces large sample sizes and frequency distributions of breeding events (Lougheed et al 2002), as did our VTG analyses. Proportions of egg-producing Marbled Murrelets. Predictions of population sizes and demographic trends must be calculated using some measure of the number of breeding individuals in the population (Ebert 1999). However, assessing the proportion of reproductive individuals in populations can be difficult, because prebreeders often disappear from the colony and nonbreeders may or may not attend (Ainley and Boekelheide 1990). The relative proportion of prebreeding and nonbreeding Marbled Murrelets in most areas is unknown, although Sealy (1975) estimated that subadults at Langara Island during the breeding season constituted ~15% of the population. By comparing radiotagged, known breeders with those predicted to breed based on presence of VTG, preliminary analyses for the Desolation Sound area in 1999 and 2000 suggest ~22% of the population are nonbreeders (Centre for Wildlife Ecology unpubl. data). In our study, the group that apparently does not produce eggs (~45%) probably includes a similar proportion of nonbreeders, with the remainder having laid eggs before or after sampling; thus, proportions of egg producers in this study (54% in 1999 and 56% in 2000) did not represent true fecundity measurements, but rather were low estimates. We suggest that variation in forest nesting habitat (e.g. in nest site availability or accessibility, or terrestrial predator density) or marine foraging areas are the main factors that influence breeding success among years. In conclusion, the use of VTG analyses as a single method permits accurate estimation of initiation, cessation, and duration of breeding stages of Marbled Murrelets. Although we found no significant annual variation in breeding chronology in Desolation Sound, we are confident that using VTG to track egg production and predict breeding chronology permits detection of interannual variation, if blood sampling begins early enough and continues throughout the entire breeding season. That method also may be used to address variation in breeding chronology across latitudes. Causes of interannual variability, along with their influence on life-history characteristics, could then be assessed. Delimiting breeding stages has implications in regulating human disturbance (e.g. logging practices and noise, Long and Ralph 1998; gill net fisheries, Carter et al. 1995) and understanding the effect of oil spills (Carter and Kuletz 1995) during the breeding season. Analysis of VTG is straightforward and can provide information about reproductive status and breeding chronology in species like the Marbled Murrelet that are difficult to study using more traditional approaches. In spite of

520 TRANQUILLA, WILLIAMS, AND COOKE [Auk, Vol. 120 the work describing their breeding chronology, the underlying reason for asynchronous breeding in Marbled Murrelets remains unclear. ACKNOWLEDGMENTS We thank the field crews who worked long, hard hours at night collecting those data. L.M.T. would especially like to thank L. and C. Lougheed for valuable advice and instruction in the field, and C. Yakel and J. Tranquilla for encouragment. Also, many thanks to B. Vanderkist and K. Salvante for lab advice and to K. Gagnon and S. Nath for lab assistance. Statistical advice was provided by B. Smith, D. Lank, and C. Schwarz, and S. Boyd and D. McFarlane made helpful comments on drafts of the manuscript. Financial support was provided by Forest Renewal British Columbia, Natural Sciences and Engineering Research Council, Centre for Wildlife Ecology, Simon Fraser University, Canadian Wildlife Service, British Columbia Ministry of Forests, Science Council of British Columbia, Timber-West Forest Ltd., International Forest Products Ltd., Western Forest Products Ltd., Weyerhauser, National Council for Air and Stream Improvement, and Pacific Forest Products Ltd. C. Smith and B. Sherman provided logistical support. LITERATURE CITED AINLEY, D. G., AND R. J. BOEKELHEIDE, EDS. 1990. Seabirds of the Farallon Islands. Ecology, Dynamics, and Structure of an Upwelling- System Community. Stanford University Press, Palo Alto, California. ASTHEIMER, L. B. 1986. Egg formation in Cassin s Auklet. Auk 103:682 693. BRADLEY, R. W., L. A. MCFARLANE TRANQUILLA, B. A. VANDERKIST, AND F. COOKE. 2002. Sex differences in nest visitation by chick-rearing Marbled Murrelets. Condor 104:180 185. CARTER, H. R. 1984. At-sea biology of the Marbled Murrelet (Brachyramphus marmoratus) in Barkley Sound, British Columbia. M.S. thesis, University of Manitoba, Winnipeg. CARTER, H. R., AND K. J. KULETZ. 1995. Mortality of Marbled Murrelets due to oil pollution in North America. Pages 261 269 in Ecology and Conservation of the Marbled Murrelet (C. J. Ralph, G. L. Hunt, Jr., M. G. Raphael, and J. P. Piatt, Eds.). U.S. Department of Agriculture, Forest Service General Technical Report PSW- GTR-152. CARTER, H. R., M. L. C. MCALLISTER, AND M. E. ISLEIB. 1995. Mortality of Marbled Murrelets in gill nets in North America. Pages 271 283 in Ecology and Conservation of the Marbled Murrelet (C. J. Ralph, G. L. Hunt, Jr., M. G. Raphael, and J. P. Piatt, Eds.). U.S. Department of Agriculture, Forest Service General Technical Report PSW-GTR-152. CHALLENGER, W. O., T. D. WILLIAMS, J. K. CHRISTIANS, AND F. VEZINA. 2001. Follicular development and plasma yolk precursor dynamics through the laying cycle in the European Starling (Sturnus vulgaris). Physiological and Biochemical Zoology 74:356 365. COOKE, F. 1999. Population studies of Marbled Murrelets (Brachyramphus marmoratus) in British Columbia. Pages 43 51 in Biology and Conservation of Forest Birds (A. W. Diamond and D. N. Nettleship, Eds.). Society of Canadian Ornithologists, Fredericton, New Brunswick. DEELEY, R. G., K. P. MULLINIX, W. WETEKAM, H. M. KRONENBERG, M. MYERS, J. D. ELDRIDGE, AND R. F. GOLDBERGER. 1975. Vitellogenin synthesis in the avian liver. Journal of Biological Chemistry 250:9060 9066. EBERT, T. A. 1999. Plant and Animal Populations: Methods in Demography. Academic Press, San Diego, California. GASTON, A. J., L. N. DE FOREST, G. DONALDSON, AND D. G. NOBLE. 1994. Population parameters of Thick-billed Murres at Coats Island, Northwest Territories, Canada. Condor 96:935 948. GASTON, A. J., AND I. L. JONES. 1998. The Auks. Oxford University Press, Oxford. GRIFFITHS, R., S. DAAN, AND C. DIJKSTRA. 1996. Sex identification in birds using two CHD genes. Proceedings of the Royal Society of London, Series B 263:1251 1256. HAMER, T. E., AND S. K. NELSON. 1995. Nesting chronology of the Marbled Murrelet. Pages 49 56 in Ecology and Conservation of the Marbled Murrelet (C. J. Ralph, G. L. Hunt, Jr., M. G. Raphael, and J. P. Piatt, Eds.). U.S. Department of Agriculture, Forest Service General Technical Report PSW-GTR-152. KAISER, G. W., A. E. DEROCHER, S. CRAWFORD, M. J. GILL, AND I. A. MANLEY. 1995. A capture technique for Marbled Murrelets in coastal inlets. Journal of Field Ornithology 66:321 333. LONG, L. L., AND C. J. RALPH. 1998. Regulation and observations of human disturbance near nesting Marbled Murrelets. U.S. Department of Agriculture, Forest Service, Redwood Sciences Laboratory, Arcata, California. LOUGHEED, C., B. A. VANDERKIST, L. W. LOUGHEED, AND F. COOKE. 2002. Techniques for investigating breeding chronology in Marbled Murrelets, Desolation Sound, British Columbia. Condor 104:319 330. MCFARLANE TRANQUILLA, L. 2001. Using multiple methods to describe breeding, stress response, and disturbance of Marbled Murrelets (Brachyramphus marmoratus). M.S. thesis, Simon Fraser University, Burnaby, British Columbia.

April 2003] Breeding Chronology in Marbled Murrelets 521 MINITAB. 2000. MINITAB Statistical Software, release 13.1. Minitab Inc., State College, Philadelphia. MITCHELL, M. A., AND A. J. CARLISLE. 1991. Plasma zinc as an index of vitellogenin production and reproductive status in the domestic fowl. Comparative Biochemistry and Physiology 100A:719 724. NELSON, S. K. 1997. Marbled Murrelet (Brachyramphus marmoratus). In The Birds of North America, no. 276 (A. Poole and F. Gills, Eds.). Academy of Natural Sciences, Philadelphia, and American Ornithologists Union, Washington, D.C. NELSON, S. K., AND T. E. HAMER. 1995. Nest success and the effects of predation on Marbled Murrelets. Pages 89 97 in Ecology and Conservation of the Marbled Murrelet (C. J. Ralph, G. L. Hunt, Jr., M. G. Raphael, and J. P. Piatt, Eds.). U.S. Department of Agriculture, Forest Service General Technical Report PSW- GTR-152. RICKER, W. E. 1984. Computation and uses of central trend lines. Canadian Journal of Zoology 62:1897 1905. SEALY, S. G. 1974. Breeding phenology and clutch size in the Marbled Murrelet. Auk 91:10 23. SEALY, S. G. 1975. Aspects of the breeding biology of the Marbled Murrelet in British Columbia. Bird-Banding 46:141 154. SPECKMAN, S. G., A. M. SPRINGER, J. F. PIATT, AND D. L. THOMAS. 2000. Temporal variability in abundance of Marbled Murrelets at sea in southwest Alaska. Waterbirds 23:364 377. SYDEMAN, W. J., J. F. PENNIMAN, T. M. PENNIMAN, P. PYLE, AND D. G. AINLEY. 1991. Breeding performance in the Western Gull: Effects of parental age, timing of breeding and year in relation to food availability. Journal of Animal Ecology 60:35 149. VANDERKIST, B. A., X. H. XUE, R. GRIFFITHS, K. MARTIN, W. BEAUCHAMP, AND T. D. WILLIAMS. 1999. Evidence of male-bias in capture samples of Marbled Murrelets from genetic studies in British Columbia. Condor 101:398 402. VANDERKIST, B. A., T. D. WILLIAMS, D. F. BERTRAM, L. LOUGHEED, AND J. P. RYDER. 2000. Indirect, physiological assessment of reproductive state and breeding chronology in free-living birds: An example in the Marbled Murrelet (Brachyramphus marmoratus). Functional Ecology 14:758 765. WANG, S., AND D. L. WILLIAMS. 1982. Biosynthesis of the vitellogenins. Journal of Biological Chemistry 257:3837 3848. WHITWORTH, D. L., J. Y. TAKEKAWA, H. R. CARTER, AND W. R. MCIVER. 1997. A night-lighting technique for at-sea capture of Xantus s Murrelets. Colonial Waterbirds 20:525 531. WIEMERSKIRCH, H. 1992. Reproductive effort in longlived birds: Age-specific patterns of condition, reproduction and survival in the Wandering Albatross. Oikos 64:464 473. ZAR, J. H. 1996. Biostatistical Analysis, 3rd ed. Prentice-Hall, Upper Saddle River, New Jersey. Associate Editor: S. Sealy