Reproductive Biology of the Brown Tree Snake, Boiga irregularis (Reptilia: Colubridae), during Colonization of Guam and Comparison with That in Their Native Range1 Julie A. Savidge2,3, Fiona J. Qualls2,4, and Gordon H. Rodda5 Abstract: Since their introduction to Guam shortly after World War II, brown tree snakes, Boiga irregularis (Merrem), have seriously impacted the biota and human population of the island. Understanding the biology of this exotic spe- cies will likely be important to the success of control programs. We compared the reproductive biology of 782 B. irregularis caught on Guam during the 1980s with results from published studies of native-range populations. Average and maximum sizes of mature snakes on Guam were larger than those from Austra- lian populations. The majority of female brown tree snakes matured at snout- vent lengths (SVLs) of 910–1,025 mm, and most males matured at SVLs of 940–1,030 mm on Guam. Based on growth rates from the early 1990s on Guam, sexual maturity is estimated to occur during a snake’s third or fourth year. Only one female (0.3%) in our data set had oviductal eggs. Clutch size was estimated at 4.3 (SD ¼ 2:2), based on large vitellogenic ovarian follicle (b30 mm in length) and oviductal egg counts. Unlike their Australian counter- parts, the Guam population reproduced year-round. Our data offer insights into the likely reproductive patterns of brown tree snakes should they infest other islands in the Pacific region. Brown tree snakes, Boiga irregularis mariannus (Wiles 1987, Wiles et al. 1995). In (Merrem), were inadvertently introduced addition, they are a recognized public health from Australasia to Guam around the end of threat (Fritts et al. 1990, 1994) and a substan- World War II, and by the mid-1980s the tial drain on the island’s economy (Fritts et al. snakes were abundant and distributed island- 1987, Fritts 2002). Accordingly, researchers wide (Savidge 1987). To date, they have been and wildlife management agencies continue responsible for the severe decline or extirpa- to develop tools aimed at controlling and tion of the island’s native forest bird species eradicating brown tree snakes from Guam (Savidge 1987), loss of two lizard species and stopping their spread to other islands. (Rodda and Fritts 1992), and a decline in Consideration of biological factors when de- numbers of the Mariana fruit bat, Pteropus m. signing eradication and control programs al- lows for maximum effectiveness with limited resources. 1 Preparation of this work was supported by the U.S. Despite being the most intensively studied Department of the Interior’s Office of Insular Affairs. Boiga species, relatively little is known about Manuscript accepted 16 June 2006. the reproductive biology of the brown tree 2 Department of Fishery and Wildlife Biology, Colo- snake, either in its native range (Indonesia to rado State University, Fort Collins, Colorado 80523. Solomon Islands and the north and east rims 3 Corresponding author: [email protected]. 4 Current address: Jones County Junior College, 900 of Australia) or on Guam. We present results South Court Street, Ellisville, Mississippi 39437. from a study of the reproductive characteris- 5 U.S. Geological Survey, Fort Collins Science Cen- tics of the Guam population during the 1980s ter, 2150 Centre Avenue, Fort Collins, Colorado 80526. and include comparisons with data collected from native-range populations (in this case, Pacific Science (2007), vol. 61, no. 2:191–199 Australia and Papua New Guinea). Combined : 2007 by University of Hawai‘i Press with data from the 1980s on food habits and All rights reserved population characteristics of brown tree snakes 191 192 PACIFIC SCIENCE . April 2007 on Guam (Savidge 1988, 1991), these data on ever, growth rates over 8.8 to 16.6 months reproductive biology provide a baseline for were available for 36 wild snakes that initially future comparisons and insight into the colo- ranged in SVL from approximately 580 to nization process by supplying a snapshot of 1,250 mm (Rodda et al. 1999). We calculated the reproductive characteristics of a recently an estimated mean age at maturity using our established brown tree snake population. calculated mean sizes for sexual maturation and mean growth rates reported in Rodda materials and methods et al. (16.8 mm/month and 13.1 mm/month for males and females, respectively), assuming Between 1982 and 1987 one of the authors growth is constant from a hatching size of ( J.A.S.) recorded morphological and repro- 375 mm SVL. ductive data from 782 brown tree snakes We used linear regression to analyze the (455 males, 327 females) that were collected relationship between SVL and testis mass. from numerous sites on Guam. Approxi- We plotted the residual scores from the re- mately 65% of snakes in the study were live- gression (i.e., measures of testis mass cor- captured by hand or by trapping in nonurban rected for the influence of body size) and locations. The remainder was obtained in ur- used analysis of variance (ANOVA) to com- ban areas, including a small number (<10%) pare the mean residual testis mass among of roadkills. After collection, live snakes were months. A square-root transformation nor- killed by freezing. Sex, snout-vent length malized the testes mass data. Statistical signif- (SVL), total length, and mass of thawed icance was set at P a :05. snakes were recorded. In female snakes, ovar- ian follicles and oviductal eggs were grouped results into size classes (<5 mm, 5–19 mm, 20–29 mm, b30 mm), the length of the largest Size and Age of Guam Brown Tree Snakes at ovarian follicle or oviductal egg measured, Maturity and the condition of oviducts recorded. We Female B. irregularis matured over a large considered females sexually mature if oviduc- range of sizes on Guam (Table 1). The small- tal eggs were present, ovarian follicles were est female to show evidence of maturation b5 mm diameter, or oviducts were thickened measured 840 mm SVL. The largest female (Shine 1987, 1991). Oviducts of 57 mature fe- to show no evidence of maturation measured males were examined for presence of sperm 1,110 mm SVL. Removal of outlying data by dissecting out the oviduct, gently washing points (the smallest 5% of mature females it with water, and examining the fluid micro- and largest 5% of immature females) indi- scopically (Shine 1977). To assess activity of cates that 90% of females matured at SVLs females with large follicles or oviductal eggs, between 910 and 1,025 mm. Males matured the stomach was examined for evidence of re- over an even greater size range. The smallest cent foraging. The length of the right testis mature male measured 805 mm SVL, and the and the combined mass of both testes were largest immature male measured 1,135 mm recorded for all males. In addition, the condi- SVL. However, 90% of males matured at tion of the vasa deferentia and the presence SVLs between 940 and 1,030 mm. of spermatozoa were recorded for 147 indi- If sexual maturation occurs at the mean viduals. The latter was determined by exam- sizes estimated here and growth is constant, ining small macerated fragments of the vasa at the mean growth rates reported in Rodda deferentia under a phase-contrast microscope et al. (1999), median age at maturity would (Shine 1977). Males were scored as mature if be 3.1 yr (males) and 3.7 yr (females). they showed fully developed convolutions (an indication of sperm presence) or spermatozoa Females with Large Follicles or Oviductal Eggs were present. We had no information on the maturation Fourteen females (4.3%) had vitellogenic fol- state of wild individuals of known age. How- licles b30 mm long, and an additional six fe- Brown Tree Snake Reproduction . Savidge et al. 193 TABLE 1 Size and Mass at Maturity of B. irregularis Collected on Guam from 1982 to 1987 and from Australian Populations (Shine 1991) Population Trait Males Females Guam SVL (mm) Sample range (n) 375–2,400 (455) 350–1,650 (327) Mature snake mean G SD (n) 1,308 G 298 (191) 1,078 G 138 (124) Mature snake extremes 805–2,400 840–1,650 Total length Sample range (n) 480–2,910 (453) 453–1,990 (326) (mm) Mature snake mean G SD (n) 1,652 G 362 (191) 1,366 G 165 (123) Mature snake extremes 1,030–2,910 1,080–1,990 Mass (g) Sample range (n) 7–2,300 (453) 7–1,800 (327) Mature snake mean G SD (n) 523 G 538 (191) 230 G 228 (124) Mature snake extremes 60–2,300 67–1,800 New South Wales, SVL (mm) Mature snake mean G SD (n) 1,015 G 249 (42) 976 G 157 (21) Australia Mature snake range 624–1,630 637–1,218 Northern Territory, SVL (mm) Mature snake mean G SD (n) 1,035 G 163 (24) 974 G 118 (22) Australia Mature snake range 850–1,535 780–1,225 Queensland, Australia SVL (mm) Mature snake mean G SD (n) 1,012 G 163 (59) 926 G 100 (38) Mature snake range 675–1,470 725–1,185 Note: SVL, snout-vent length; SD, standard deviation; n, sample size. males had follicles between 20 and 29 mm follicles b5 mm long or oviductal eggs were long. Only one female (0.3%) had oviductal captured in every month (Figure 1). Mature eggs. Of females with follicles b30 mm long, male brown tree snakes exhibited a positive 79% had prey items in their stomachs. The linear relationship between SVL and the one female with oviductal eggs had no prey square root of testis mass (r 2 ¼ 0:75; F ¼ items in its stomach.