Western North American Naturalist

Volume 72 Number 3 Article 10

11-5-2012

Population abundance and sex ratio of the viviparous freshwater baenschi () throughout its range in western Mexico

J. Jaime Zúñiga-Vega Universidad Nacional Autónoma de México, México, [email protected]

Ana L. Hernández-Rosas Universidad Nacional Autónoma de México, México, [email protected]

Alejandro Molina-Moctezuma Universidad Nacional Autónoma de México, México, [email protected]

Hibraim A. Pérez-Mendoza Universidad Nacional Autónoma de México, México, [email protected]

Fernanda R. Rodríguez-Reyes Universidad Nacional Autónoma de México, México, [email protected]

Follow this and additional works at: https://scholarsarchive.byu.edu/wnan See next page for additional authors Part of the Anatomy Commons, Botany Commons, Physiology Commons, and the Zoology Commons

Recommended Citation Zúñiga-Vega, J. Jaime; Hernández-Rosas, Ana L.; Molina-Moctezuma, Alejandro; Pérez-Mendoza, Hibraim A.; Rodríguez-Reyes, Fernanda R.; Bravo-Espinosa, Yolotzin M.; and Espinosa-Pérez, Héctor (2012) "Population abundance and sex ratio of the viviparous freshwater fish Poeciliopsis baenschi (Poeciliidae) throughout its range in western Mexico," Western North American Naturalist: Vol. 72 : No. 3 , Article 10. Available at: https://scholarsarchive.byu.edu/wnan/vol72/iss3/10

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Western North American Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Population abundance and sex ratio of the viviparous freshwater fish Poeciliopsis baenschi (Poeciliidae) throughout its range in western Mexico

Authors J. Jaime Zúñiga-Vega, Ana L. Hernández-Rosas, Alejandro Molina-Moctezuma, Hibraim A. Pérez- Mendoza, Fernanda R. Rodríguez-Reyes, Yolotzin M. Bravo-Espinosa, and Héctor Espinosa-Pérez

This article is available in Western North American Naturalist: https://scholarsarchive.byu.edu/wnan/vol72/iss3/10 Western North American Naturalist 72(3), © 2012, pp. 357–368

POPULATION ABUNDANCE AND SEX RATIO OF THE VIVIPAROUS FRESHWATER FISH POECILIOPSIS BAENSCHI (POECILIIDAE) THROUGHOUT ITS RANGE IN WESTERN MEXICO

J. Jaime Zúñiga-Vega1,3, Ana L. Hernández-Rosas1, Alejandro Molina-Moctezuma1, Hibraim A. Pérez-Mendoza1, Fernanda R. Rodríguez-Reyes1, Yolotzin M. Bravo-Espinosa1, and Héctor Espinosa-Pérez2

ABSTRACT.—We estimated population abundance and sex ratio for a freshwater live-bearing fish species (Poeciliopsis baenschi) that inhabits fluvial systems in western Mexico. We conducted mark-recapture experiments in 6 distinct popu- lations. We used both our catch data as well as estimates of capture probability to calculate the number of individuals per unit water volume for adult males, adult females, and juveniles. Overall, abundance varied among sites from 23.8 individuals ⋅ m–3 to 363.2 individuals ⋅ m–3. In most of our study sites, the estimated sex ratio was markedly biased toward females (up to 11 females per male). We discuss the potential causes and implications of the observed demo- graphic features.

RESUMEN.—Estimamos abundancia poblacional y proporción sexual para una especie de pez vivíparo de agua dulce (Poeciliopsis baenschi) que habita sistemas fluviales en el occidente de México. Llevamos a cabo experimentos de mar- caje y recaptura en seis poblaciones distintas. Utilizamos tanto nuestras capturas observadas así como estimaciones de probabilidad de captura para calcular el número de individuos por unidad de volumen de agua para machos adultos, hembras adultas y juveniles. En general, la abundancia varió entre sitios desde 23.8 individuos por m3 y hasta 363.2 individuos por m3. En la mayoría de nuestros sitios de estudio, la proporción sexual estimada estuvo marcadamente ses- gada hacia las hembras (hasta 11 hembras por cada macho). Discutimos las causas potenciales y las implicaciones de los rasgos demográficos observados.

Population abundance and sex ratio are management plans (Caswell 2000, 2001, Wil - basic demographic parameters that still need liams et al. 2002). description for a large number of Viviparous of the family Poeciliidae species. In the Neotropics, studies estimating are particularly interesting because most spe - demographic parameters of freshwater fishes cies are restricted to tropical freshwater envi- are notably scarce in comparison with those ronments and several have markedly narrow focused on Nearctic and Palearctic regions distributional ranges (Evans et al. 2011). The (e.g., Killgore et al. 2007, High et al. 2008, high level of endemism observed in many of Young et al. 2010). As a consequence, little is these fishes makes them vulnerable to human- known about the relative abundance, popula- caused disturbance (Duncan and Lockwood tion characteristics (e.g., sex ratio), and persis- 2001, Contreras-Balderas et al. 2003). Unfor- tence probabilities of these organisms in fluvial tunately, only a few widespread species have systems of the Neotropics (but see Arce-Uribe been demographically characterized (e.g., 2006, Ramos-Santiago et al. 2006, Canto-Maza Brachyrhaphis rhabdophora, Gambusia affinis, and Vega-Cendejas 2007). The current rate of Gambusia holbrooki, Heterandria formosa, and change in land use and water pollution in Poecilia reticulata; Jordan et al. 1998, Bronikow - highly diverse neotropical ecosystems (Miguel ski et al. 2002, Smith 2005, Richardson et al. 1991, Achard et al. 2002) also makes it impera- 2006, Johnson and Zúñiga-Vega 2009, Johnson tive to know basic demographic traits of fresh- and Bagley 2011), and we know little about basic water fish species. This demographic informa- population traits such as abundance and sex tion can then be used to guide conservation or ratio for poeciliids with restricted geographic

1Departamento de Ecología y Recursos Naturales, Facultad de Ciencias, Universidad Nacional Autónoma de México. Ciudad Universitaria 04510, Distrito Federal, México. 2Colección Ictiológica, Departamento de Zoología, Instituto de Biología, Universidad Nacional Autónoma de México. Ciudad Universitaria 04510, Distrito Federal, México. 3E-mail: [email protected]

357 358 WESTERN NORTH AMERICAN NATURALIST [Volume 72



Fig. 1. Geographic location of the 6 studied populations of Poeciliopsis baenschi. Numbers indicate study sites as per Table 1. distributions. In general, those poeciliid species larger than adult males: 24.4 and 21.3 mm whose population abundances have been de - mean standard length (SL), respectively. How- scribed are usually dominant species in their ever, both sexes appear to mature at similar habitats, with population densities varying sizes: 18.6 and 19.5 mm SL for females and between 8 and 44 individuals per m2 (Meffe males, respectively. Females exhibit superfeta- and Snelson 1989, Jordan et al. 1998, Reznick tion, which is the ability to simultaneously et al. 2001, Alexandre et al. 2010). More inter- carry multiple broods of embryos at different esting is the fact that the number of individu- stages of development (Turner 1937, Scott and als can vary drastically among populations, Johnson 2010). To date, nothing is known and factors such as food and shelter availabil- about the behavior, physiology, feeding habits, ity, or the presence or absence of predators or population dynamics of this viviparous fish and competitors, have been proposed as the (Miller et al. 2005). main causes of this spatial variation (Chapman In this study, we estimated population abun- et al. 1991, Belk and Lydeard 1994, Reznick et dance and sex ratio for 6 different populations al. 2001, Richardson et al. 2006). Neverthe- of P. b a e n s c h i (Fig. 1). The selected popula- less, a complete picture of the fundamental tions span most of the geographic distribution demographic parameters of this fish family of the species. We selected P. baenschi as our requires the study of species with more re- focal species for 2 main reasons. First, its popu- stricted geographic distributions. lations occur only in a restricted area of the Poeciliopsis baenschi is one example of a Pacific Slope in western Mexico (Miller et al. viviparous fish species inhabiting only a few 2005). In this region, several factors such as river drainages in a neotropical dry forest. This water pollution, urban development, and exten - species is endemic to streams and rivers in the sive agriculture are negatively affecting the states of Jalisco and Colima in western Mexico main fluvial systems and, in consequence, are (Miller et al. 2005). Like other species in the threatening the native species (Ceballos and family Poeciliidae, males internally inseminate García 1995, Castillo et al. 2009, Masek et al. females using a modified anal fin (gonopo - 2011). Second, no study has been conducted dium), and females give birth to free-swimming on the demographic parameters of this species, young. This species exhibits marked sexual and thus, no quantitative information is avail- dimorphism in size, and adult females are able to guide future conservation actions. 2012] ABUNDANCE OF POECILIOPSIS BAENSCHI 359

Essentially, we aimed to answer this question: induced disturbance, such as close proximity what is the influence of different rivers on the to rural human settlements and the occurrence population abundance and sex ratio of P. baen- of waste materials in the riverbed. Based on schi? Given that the 6 studied rivers differ these criteria, we classified the sites as either markedly in key environmental conditions, disturbed or pristine. Table 1 shows a descrip- such as the presence or absence of potential tion of the study sites, as well as sample sizes predators or competitors, we expected to find (i.e., number of captured and marked fish per marked differences among populations in both population). abundance and sex ratio. Our results repre- Fish Sampling sent a significant contribution to the knowl- edge of a freshwater fish species endemic to To estimate population abundance and sex the tropical dry forest of western Mexico and a ratio we conducted capture-mark-recapture reference for future similar studies quantify- experiments during the dry seasons of 2009 ing basic demographic traits of neotropical and 2010. We visited the sampling sites once a fauna. week during a 5-week period. The only excep- tion was population 4 that was visited weekly METHODS during a 4-week period. Sites 1–4 were sam- pled during October–November 2009. Sites 5 Study Sites and 6 were sampled during October–November For this study we selected 6 different locali- 2010. At each population we selected one focal ties that represent independent fluvial sys- pool, and on each visit we attempted to collect tems with no inland connection among them all fish in the pool by repeated seining. For this (Fig. 1). Thus, our study localities represent purpose, we used handheld seine nets (1.3 m different populations of P. baenschi. These sites high × 5 m long, 8-mm mesh size). Upon first span a large proportion of the geographic range capture, each fish was anesthetized using MS- in which P. b a e n s c h i can be found (Miller et al. 222, measured, and marked individually with 2005). In fact, we selected these sampling visible implant elastomer (VIE) tags (North- sites because they correspond to most of the west Marine Technology, Inc.) injected in the river drainages which this species inhabits. caudal peduncle. Also on each first capture, The only exception was a small stream located we determined the sex of each individual by 50 km south of Puerto Vallarta city, which is the presence or absence of the male gono- separated from our closest sampling site by podium. On subsequent recaptures the indi- about 110 km. For logistic reasons, we could vidual mark was identified and individuals not sample at this relatively distant location. were measured again. On every visit we caught Thus, our study sites span most of the types of both previously marked and unmarked fish rivers and streams in which P. b a e n s c h i occurs. (except for the first occasion in which only The sampling rivers differ markedly in cer- unmarked individuals were caught; Table 2). tain environmental variables (Table 1). First, All the individuals caught without a mark were the particular freshwater fish species with individually marked as well. We only marked which P. b a e n s c h i coexist vary among sites and followed fish larger than 15 mm SL, (Miller et al. 2005, Álvarez-Romero et al. 2008). because smaller individuals were seriously Interestingly, in 4 of the 6 sampling localities, affected by the anesthetic and the injection of members of the family Cichlidae are present, VIE tags. Our weekly sampling procedure and these fishes might prey on P. b a e n s c h i allowed us to obtain individualized recapture (Reznick et al. 1996, Johnson and Belk 2001). (encounter) histories for each marked fish Second, 2 sampling localities were small throughout a 5-week period (4-week period in streams, and the other 4 were relatively large the case of site 4). rivers. This dichotomous classification of our Data Analyses sites as either stream or river depended on the width of the riverbed. We arbitrarily defined To estimate population abundance we used rivers as waterways wider than 5 m and streams both our catch data and estimates of capture as those narrower than 5 m. Third, our sam- probability. We could not consider the number pling localities also differed in the presence of individuals caught per population or sampling (or absence) of conspicuous signs of human- occasion as direct estimates of population 360 WESTERN NORTH AMERICAN NATURALIST [Volume 72 StreamRiver Pristine Stream Pristine River Disturbed River Disturbed Pristine River Pristine a a b b b b b c c c,d c,d c,d c,d a,d a,d b b b a a a a a a ed disturbance, such as close proximity to rural human ed disturbance, such as close proximity es throughout the text. ), total water volume that was sampled on every visit (i.e., the esti- n Poecilia butleri Poecilia Poecilia butleri Poecilia resolanae Xenotaenia butleri Poecilia Ilyodon furcidens resolanae Xenotaenia Oreochromis aureus butleri Poecilia butleri Poecilia resolanae Xenotaenia Oreochromis aureus butleri Poecilia reticulata Poecilia turrubarensis Poeciliopsis Allodontichthys zonistius Ilyodon furcidens Cichlasoma istlanum Oreochromis aureus reticulata Poecilia turrubarensis Poeciliopsis Allodontichthys zonistius Ilyodon furcidens Cichlasoma istlanum Oreochromis aureus ) Co-occuring species River/stream Pristine/disturbed 3 volume (m Sampled water We considered rivers as waterways with a bed wider than 5 m and streams those We n P. baenschi. P. W 573 4.0 WW 885W 851 40.6 W 929 2.48 W 977 7.13 530 13.5 16.1 33.1 18.7 20.3 32.0 39.8 49.7 ° ° ° ° ° ° We show total number of captured and marked fish (Total show total number of captured and marked fish (Total We N, 104 N, 104 N, 104 N, 104 N, 103 N, N, 103 N, 44.7 29.5 12.1 23.4 54.6 10.4 ° ° ° ° ° ° Poeciliopsis baenschi. Poeciliopsis 1. Studied populations of ABLE T Exotic species Exotic Family Goodeidae Family Family Cichlidae Family Family Poeciliidae Family mated water volume of each sampling pool), and species co-occurring with classified sites as either disturbed or pristine based on the presence absence of signs human-induc narrower than 5 m. We settlements and the occurrence of waste materials in riverbed. Numbers assigned to sites are used for identification purpos 4Río Pilastón19 12 Upper Río Purificación Cuzalapa Río 3 19 4Río Chandiablo Arroyo 5 19 Coahuayana Río 19 6 18 Armería Río a b c d 19 Site Name Geographic location Total 2012] ABUNDANCE OF POECILIOPSIS BAENSCHI 361 We 4.7) 18.8) 24.0) 49.5) 67.0) 82.6) 86.8) 417.8) 167.1) 157.6) 22.2) 197.6) – – – – – – – – – – – – n 5 at site 4. Poeciliopsis baenschi. Poeciliopsis 6.5) (1) (0.01 79.2) (34) (40.2 5.7)19.8) (7) (38)34.8) (46) (0.6 (10.4 (19.5 38.5)471.3) (1) (29) (29.4 (111.2 55.2) 146.9) — — — 202.1) (22) (72.8 46.6)137.2) (3) (45) (13.7 (82.9 151.4) (6)14.7) (49.2 (8) (9.2 – – – – – – – – corresponds to sample size in the first capture – – – – – – n nitial 6.6) (11)21.9)22.0) (41) (0.7 34.1) (3) (55)317.3) (10.8 (19) (0.01 (16.4 (83.4 34.2)612.8) (2) (27) (11.8 (152.9 37.9) (1) (0.01 103.8) (47) (47.3 164.5) (18) (22.8 38.8)144.0) (6) (61) (8.5 (68.3 310.6) (6)21.3) (44.2 (8) (6.2 – – – – – – – – – – – – – – )/water estimated abun- volume of the corresponding sampling pool] in order to express p ) per sex/stage class and sampling occasion for 6 populations of ) per sex/stage / N n (95%) are shown in parentheses. Description of sites is given in Table 1. (95%) are shown in parentheses. Description of sites is given Table N was calculated as [( N 5.5)25.4) (9)11.6) (44)36.2) (3) (56)256.8) (0.9 (12.4 (18) (0.01 (18.2 (137.4 75.3)207.6) (8) (69) (6.8 (81.9 398.5)51.1) (13)614.4) (2) (33)24.1) (96.7 (8) (11.6 (247.4 (9.1 127.0) (53) (63.7 154.5) (39) (45.4 18.6) (2) (0.01 – – – – – – – – – – – – – – ), and estimated number of individuals ( p Occasion 2 Occasion 3 Occasion 4 Occasion 5 (7)(0) (0.7 (9) (0.01 (111.4 (3) (13.7 (1)(3) (17.5 (10.3 (3) (0.5 . Water volume of the corresponding sampling pools is given in Table 1. Notice that no sampling effort was conducted on occasio volume of the corresponding sampling pools is given in Table . Water (28)(35) (14.3 (18.3 (46) (115.9 (20)(30) (126.1 (276.7 (40) (77.3 (22) (51.3 ______3 nNp npNnpN nnpNnpN were only available for occasions 2–5. N ), capture probability ( Initial n and p 2. Sample size ( ABLE T MalesFemalesJuveniles classesAll sex/stage 37 227 150Males 40 39 230 172 0.31 0.21 0.21 19 0.09 27.3 3.1 19.8 112 5.4 84 232 149 47 0.18 0.22 0.21Juveniles 0.31 36 184.1 classesAll sex/stage 26.1 17.1 0.09 3.7 323 104 10.2 63 182 131 271 0.18 40 0.24 0.17 0.21 60 0.31 227.4 11 0.19 161.7 25.6 15.3 0.09 3.2 44.5 65 207 3.1 196 124 0.18 0.26 0.14 0.21 32 31 142.7 113.0 0.30 34.5 40 0.19 14.6 0.09 2.6 56 22.8 167 11.3 0.19 0.23 37 120.0 102.7 0.19 182 27.6 0.18 140.2 63 0.18 48.1 FemalesJuveniles classesAll sex/stage 337 171Males 54 200 101Females 0.18 0.15 15 0.18 445.5 262.3 41 219Males 34.3 192 28 78 183Females 0.18 0.23 0.15 0.25 10 430.1 203.6 17.2 102.1 0.18 63 261 22.9 113 14 151 37 193 25 0.15 0.25 0.11 0.14 0.23 0.15 11 312.1 83.7 102.9 97.8 9.5 15.2 0.18 122 25.1 113 221 40 17 0.19 31 0.16 0.25 0.16 0.12 0.23 264.5 26 104.9 63.3 103.4 18.4 10.4 0.18 124 58.1 94 23 25 0.11 0.25 0.22 0.08 84.9 53.6 15.7 22.0 — — — dance as number of individuals per m also show data for all sex/stage classes pooled together (overall population abundance per and sampling occasion). I also show data for all sex/stage occasion. Estimates of Number of previously marked individuals (recaptures) and confidence intervals for Site 1 Site 2 Site 3 Site 4 362 WESTERN NORTH AMERICAN NATURALIST [Volume 72 5.6) 37.6) 343.9) 181.4) 41.9) 48.2) 11.2) 33.7) – – – – – – – – 41.1) (38)63.1) (2) (14.9 (3.3 30.3) — — — 293.3) (23) (88.4 195.3)57.9) (16) (5) (50.2 (6.5 15.8) (35) (6.9 38.5) (1) (3.1 134.7) —11.0) — (2) — (0.03 – – – – – – – – – – 35.0) (44) (18.9 17.4) (36) (9.6 30.3) (1) (0.01 280.5) (21) (78.5 151.4)28.3) (14) (2) (67.9 (4.0 21.2) (5) (0.01 20.6) (6) (5.6 197.9) (20)10.2) (2) (61.1 (0.3 – – – – – – – – – – 76.6) (10) (47.1 29.2) (43) (17.8 18.7) (36) (10.7 8.5) (6) (0.4 29.8) (0) (0.01 112.3) (12) (59.9 63.6) (0) (3.9 23.7) (2) (2.7 27.3) (1) (3.1 185.4) (41) (52.7 – – – – – – – – – – Occasion 2 Occasion 3 Occasion 4 Occasion 5 (3) (5.4 (0) (3.2 (5) (2.4 (3) (0.2 (1) (0.01 (18) (40.7 (15) (13.4 (52) (17.3 (44) (11.7 (26) (64.1 ______nNp npNnpN nnpNnpN Initial 2. Continued. ABLE Juveniles classesAll sex/stage 144 75 60 0.20 17 0.12 76.5 34.5 84 0.12 8 0.12 170.2 16.1 128 0.17 17 185.9 0.14 30.9 117 0.14 14 216.1 0.14 24.2 MalesFemales 17 52 8 35 0.15 0.19 13.4 45.0 7 69 0.15 0.17 11.9 99.2 20 91 0.18 0.17 27.7 131.6 85 18 0.18 0.18 115.8 25.7 Juveniles classesAll sex/stage 223 43 123 0.33 32 0.13 23.2 14.8 110 0.26 24 0.13 26.4 11.8 125 0.26 46 30.0 0.13 22.0 90 0.21 36 26.3 0.12 18.4 MalesFemales 20 160 10 81 0.14 0.33 4.4 15.2 12 74 0.14 0.33 5.3 14.0 67 12 0.33 0.13 12.7 5.6 48 0.33 6 0.13 9.1 2.8 All sex/stage classesAll sex/stage 375 224 0.13 124.8 269 0.16 125.3 193 0.15 97.9 T Juveniles 51 17 0.12 10.4 17 0.12 10.5 17 0.12 10.5 Site 6 Site 5 2012] ABUNDANCE OF POECILIOPSIS BAENSCHI 363 abundance because we might have missed a lation abundance (along with SE) has been large number of individuals actually present implemented in previous studies on demo- during our sampling events. Therefore, we graphic parameters of animal populations (e.g., began by estimating capture probability (p) Wood et al. 1998, Armstrong et al. 2005, Zúñiga- per sex, stage class, and sampling occasion for Vega 2011). For each population, we reported each study site. We structured the populations weekly estimates for number of individuals + into 3 groups or sex/stage classes: (1) adult per sex/stage class (Ni – SE and 95% confi- males: fishes with gonopodium present and dence intervals). larger than 19.5 mm SL; (2) adult females: In addition, we calculated overall popula- fishes with gonopodium absent and larger than tion abundance (juveniles, males, and females 18.6 mm SL; and (3) juveniles: fishes larger together) per sampling occasion as the ratio than 15 mm SL and smaller than 19.5 and 18.6 between the total number of captured fish per mm SL for males and females, respectively occasion (total n) and the estimated occasion- (Scott and Johnson 2010). As sex is not evident specific values of p for all the individuals in small fishes (<18.6 mm SL) because of pooled into a single group. Standard errors for incomplete development of the gonopodium, these estimates of overall population abun- we did not distinguish males from females in dance were also calculated as explained above. the juvenile class. Sex ratio was simply calculated as the ratio We used maximum likelihood procedures between the estimated number of adult males implemented in the program MARK to esti- and females present at each population. For mate p for each sex/stage class, sampling occa- comparative purposes, all estimates are given sion, and population (White and Burnham as per unit water volume (individuals ⋅ m–3), 1999). MARK finds the values of p that maxi- because the volume of the selected sampling mize the likelihood of obtaining the observed pool was different among study sites (Table 1). encounter histories. Additionally, this com- puter program uses likelihood functions to cal- RESULTS culate standard errors for p (Lebreton et al. 1992). Estimates of p were only available for The estimated number of individuals in occasions 2–5 given that there were no previ- each sex/stage class per sampling occasion and ously marked during the first sam- per unit water volume (m3) was widely vari- pling occasion. able among populations (Fig. 2). The esti- Using estimates of p, we calculated the mated number of males per m3 (+–SE) was as number of individuals present per sex/stage low as 2.6 +– 1.0 in site 1 and as high as 227.4 + class, sampling occasion, and population (Ni) – 45.9 in site 2. The estimated number of as follows: females per m3 was as low as 9.1 +– 1.1 in site 5 and as high as 262.3 +– 69.5 in site 2. The 3 Ni = ni / pi , estimated number of juveniles per m was as low as 3.1 +– 1.7 in site 1 and as high as 58.1 +– where n represents the number of individuals 14.6 in site 2. Table 2 shows all estimates for caught (sample size) and i represents any com- number of individuals per sex/stage class, sam- bination of sex/stage class, sampling occasion, pling occasion, and population (expressed per and population. Notice that this formula ac - m3 +– 95% confidence intervals) as well as counts for the specific probability of capturing sample sizes (ni) and capture probabilities ( pi). animals when estimating the actual number of The calculated overall population abun- individuals present. Estimates of N could only dance per sampling occasion was also spatially be calculated for occasions 2–5 because no variable. Site 2 exhibited the highest popula- estimate of p was available for occasion one. tion abundances of P. b a e n s c h i (between 312.1 Standard errors (SE) for Ni were calculated as and 445.5 individuals ⋅ m–3; Fig. 2, Table 2). The lowest population abundances occurred 2 SE(Ni) = (ni × SE[pi]) / pi . at site 5 (between 23.2 and 30.0 individuals ⋅ m–3; Fig. 2, Table 2). Overall population abun- Confidence intervals (95%) were constructed dances (averaged across sampling occasions) as 1.96 × SE(Ni), assuming a normal distribu- were as follows: site 1: 25.5 +– 3.1; site 2: 363.2 +– tion of Ni. This procedure for estimating popu- 60.5; site 3: 119.3 +– 12.8; site 4: 119.3 +– 28.6; 364 WESTERN NORTH AMERICAN NATURALIST [Volume 72

        

          

 

  

 

            

                 

         

 

 

 

 

             

                

          

 

  

 

            

                 

Fig. 2. Estimated number of individuals per sex/stage class, population, and sampling occasion. Estimates are shown per unit water volume (m3). For visualization purposes, the scales of the y-axes differ depending on the abundance of individuals per population. Estimates were available beginning with sampling occasion 2. Site 4 was visited only in 4 occasions instead of 5. Vertical bars denote one standard error around the mean. 2012] ABUNDANCE OF POECILIOPSIS BAENSCHI 365 site 5: 23.8 +– 2.7; and site 6: 148.1 +– 29.6 parasites, must be measured with more preci- individuals ⋅ m–3. sion in order to understand the factors causing A recurring pattern (present in 5 of our 6 the observed variation among populations in study sites) that arose in the population esti- the abundance of P. b a e n s c h i (Belk and Lydeard mates was that sex ratio was markedly biased 1994, Garcia et al. 2004, Richardson et al. toward females (Fig. 2). The largest difference 2006, Alexandre et al. 2010). in abundance between sexes occurred at site 6. Careful attention should be paid to those There the maximum estimate for number of sites at which the abundance of P. baenschi is females per m3 was 131.6 +– 32.5, whereas the remarkably low (i.e., sites 1 and 5). This is the minimum estimate for number of males per first study measuring the number of individu- m3 was only 11.9 +– 4.7 (Table 2); this repre- als per m3 for this species, and thus, we do not sents a sex ratio (females : males) equal to 11.1:1. know whether the observed abundances were In fact, even juveniles were more abundant higher in the past. If so, 2 hypotheses that than adult males in sites 3 and 5 (Fig. 2). The deserve further research arise. The first is that average sex ratios (females : males) across sam- population size at these localities is decreasing pling occasions for study sites 1, 3, 4, 5, and 6 and that there is now certain risk of local were 6.0:1, 4.9:1, 3.1:1, 2.6:1, and 4.4:1, extinction. If this were the case, we would rec- respectively. The only exception from this ommend the implementation of conservation- female-biased sex ratio was site 2, in which sex based research programs at these 2 localities. ratio was approximately 1:1 (Fig. 2, Table 2). In fact, other poeciliids are now considered locally extinct, presumably as a result of recent DISCUSSION human-induced disturbance (e.g., Priapella bonita and Xiphophorus couchianus; Contreras- Population Abundance Balderas et al. 2003). The second hypothesis is We found remarkable spatial variation in that there is year-to-year natural variability in population abundance of P. b a e n s c h i . Given the population abundance of this species. Pre- the contrasting ecological conditions observed vious studies have documented temporal vari- in the 6 studied rivers, this result is not sur- ability in demographic parameters of poeciliid prising. However, the site-specific environ- fishes (e.g., Gómez-Márquez et al. 2008, Súarez mental variables that we recorded cannot et al. 2009). We could have sampled sites 1 account for the differences among populations and 5 during years in which environmental in the abundance of P. b a e n s c h i . This species conditions promoted low abundance of P. can be either abundant or relatively rare in baenschi. In fact, we recognize that our mark- the presence of the same array of potential recapture experiments were conducted in dif- predators and competitors. For instance, in ferent years (2009 for sites 1–4 and 2010 for sites 5 and 6, P. b a e n s c h i coexists with the sites 5 and 6). Thus, an important assumption same members of the families Poeciliidae, underlying our results and conclusions is that Goodeidae (potential competitors), and Cichli- interannual variation in the demographic char- dae (potential predators; Table 1). However, in acteristics of P. b a e n s c h i is minimal in compari- site 5, P. baenschi is relatively rare, whereas in son with variation among populations. An site 6 it is considerably abundant (Fig. 2). The alternative hypothesis that also remains to be size of the riverbed could not determine the tested is that the abundance of P. b a e n s c h i is abundance of P. b a e n s c h i either. These fish can naturally low at sites 1 and 5. be abundant in relatively large rivers, such as Sex Ratio that sampled in site 2, as well as in small streams, such as that sampled in site 3. In con- Our most remarkable finding was the sex trast, P. baenschi abundance can be markedly ratio of P. b a e n s c h i . In most of our study sites, low in both rivers (e.g., site 5) and streams females were markedly more abundant than (e.g., site 1). The same was true for human- males. Sex ratio can be as biased as 11 females induced disturbance; P. b a e n s c h i was abun- per male. Biased sex ratios have been docu- dant in both pristine and disturbed river mented previously in freshwater fishes, includ- drainages, such as sites 2 and 4, respectively. ing poeciliids (Hughes 1985a, 1985b, Zulian et Certainly, these and other ecological factors, al. 1995, Cognato and Fialho 2006, Andrade et such as food availability or the presence of al. 2008). In contrast, in the population where 366 WESTERN NORTH AMERICAN NATURALIST [Volume 72 we observed the highest abundance (site 2), light 2 main findings. First, population abun- sex ratio was 1:1. This interesting difference dance varies widely among populations. Sec- may have important consequences in terms of ond, sex ratio in most of the studied sites is male–male competition, sexual selection mech - noteworthy: females are considerably more anisms, and phenotypic divergence in this site abundant than males. The steps to follow should with respect to all other populations (Macías- be to calculate population growth rates and to Garcia 1994, Smith 2007). Nevertheless, the examine temporal (e.g., seasonal) variation in observed differences among populations in sex the observed demographic parameters. In fact, ratio can also result from seasonal or interan- we recognize that our data are restricted to nual variability in the number of males and the dry season, whereas both abundance and females present, such as has been reported for sex ratio can vary markedly throughout the other poeciliid fishes (Gómez-Márquez et al. year (Gómez-Márquez et al. 2008, Súarez et 2008, Súarez et al. 2009). al. 2009). What causes female-biased sex ratios? Biased sex ratios are the result of processes ACKNOWLEDGMENTS occurring during fertilization (e.g., sex-deter- mining chromosomes), during embryonic devel - We thank Josh Rasmussen for helpful opment and birth, during the juvenile stage, comments on the manuscript. Field assistance or after sexual maturity (McKellar et al. 2009). was provided by Pedro Mendoza-Hernández, Neither sex-linked inheritance nor sex chro- Claudia Molina-Zuluaga, Claudia Olivera- mosomes have been found in Poeciliopsis Tlahuel, Jonathan Maceda-Cruz, and Alejan- species (Sullivan and Schultz 1986, Schultheis dra Martínez-Blancas. Jaime Zúñiga-Gutiérrez et al. 2009). Thus, any process during fertiliza- provided logistic support. Funding for this proj- tion can be discarded for P. b a e n s c h i . Among ect was provided by the Dirección General the factors that may affect the proportion of de Asuntos del Personal Académico–Universi- sexes during embryonic development, tempera- dad Nacional Autónoma de México through ture exerts a significant effect in poeciliids, the project PAPIIT IN206309-3. Fieldwork with low water temperature resulting in more was conducted under permit no. FAUT-0240 females and high water temperature resulting issued by the Secretaría de Medio Ambiente y in more males (Ospina-Álvarez and Piferrer Recursos Naturales–México. 2008). If the proportion of sexes is 1:1 at birth in P. LITERATURE CITED baenschi, then males and females must experi- ence distinct selective pressures before and/or ACHARD, F., H.D. EVA, H.J. STIBIG, P. MAYAUX, J. GALLEGO, T. R ICHARDS, AND J.P. MALINGREAU. 2002. Determi- after maturity. This differential selection can nation of deforestation rates of the world’s humid result in biased adult sex ratios. Among the tropical forests. Science 297:999–1002. factors that can differentially affect males and ALEXANDRE, C.V., K.E. ESTEVES, AND M.A.M.D.E. females are sex-specific migration patterns, MELLO. 2010. Analysis of fish communities along a rural-urban gradient in a neotropical stream (Piraci- distinct mortality rates associated with differ- caba River Basin, São Paulo, Brazil). Hydrobiologia ential susceptibility to predators, intersexual 641:97–114. differences in resource acquisition under low- ÁLVAREZ-ROMERO, J.G., R.A. MEDELLÍN, A. OLIVERAS DE resource levels, inter- or intrasexual aggression, ITA, H. GÓMEZ DE SILVA, AND O. SÁNCHEZ. 2008. Animales exóticos en México: una amenaza para la and sex-specific reproductive costs (Girondot biodiversidad. Comisión Nacional para el Cono - and Pieau 1993, Ghalambor et al. 2004, McKel- cimiento y Uso de la Biodiversidad, Universidad lar et al. 2009). These and other factors might Nacional Autónoma de México, Secretaria de Medio interact to produce the markedly female- Ambiente y Recursos Naturales, Mexico City, Mexico. biased sex ratios observed in P. b a e n s c h i . Their ANDRADE, V.X.L., F.F.S. CAMPOS, F. LANGEANI, AND E. ROMAGOSA. 2008. Reproductive dynamics of the potential effects represent a fertile field for main species of fish in the municipal reservoir of São future research. José Do Rio Preto. Boletim do Instituto de Pesca 34:365–373. Concluding Remarks ARCE-URIBE, E. 2006. Abundancia y distribución pobla- We have provided basic demographic infor- cional de Poblana alchichica (Pisces: Atherinopsidae) charal endémico del Lago Alchichica, Puebla. Mas- mation for several populations of the livebear- ter’s thesis, Universidad Nacional Autónoma de ing fish Poeciliopsis baenschi. We can high- México, Mexico City, Mexico. 2012] ABUNDANCE OF POECILIOPSIS BAENSCHI 367

ARMSTRONG, D.P., R.S. DAVIDSON, J.K. PERROTT, J. ROY- GÓMEZ-MÁRQUEZ, J.L., B. PEÑA-MENDOZA, I.H. SALGADO- GARD, AND L. BUCHANAN. 2005. Density-dependent UGARTE, A.K. SÁNCHEZ-HERRERA, AND L. SASTRÉ- population growth in a reintroduced population of BAEZ. 2008. Reproduction of the fish Poeciliopsis North Island saddlebacks. Journal of Animal Ecology gracilis (: Poeciliidae) in Coate - 74:160–170. telco, a tropical shallow lake in Mexico. Revista de BELK, M.C., AND C. LYDEARD. 1994. Effect of Gambusia Biología Tropical 56:1801–1812. holbrooki on a similar-sized, syntopic poeciliid, Heter- HIGH, B., K.A. MEYER, D.J. SCHILL, AND E.R.J. MAMER. andria formosa: competitor or predator? Copeia 2008. Distribution, abundance, and population trends 1994:296–302. of Bull Trout in Idaho. North American Journal of BRONIKOWSKI, A.M., M.E. CLARK, F.H. RODD, AND D.N. Fisheries Management 28:1687–1701. REZNICK. 2002. Population-dynamic consequences HUGHES, A.L. 1985a. Seasonal changes in fecundity and of predator-induced life history variation in the size at first reproduction in an Indiana population of guppy (Poecilia reticulata). Ecology 83:2194–2204. the mosquitofish Gambusia affinis. American Mid- CANTO-MAZA, W.G., AND M.E. VEGA-CENDEJAS. 2007. land Naturalist 114:30–36. Distribución, abundancia y preferencias alimenticias ______. 1985b. Seasonal trends in body size of adult male del pez sapo Opsanus phobetron (Batrachoididae) en mosquitofish, Gambusia affinis, with evidence for la laguna costera de Chelem, Yucatán, México. their social control. Environmental Biology of Fishes Revista de Biología Tropical 55:979–988. 14:251–258. CASTILLO, A., C. GODÍNEZ, N. SCHROEDER, C. GALICIA, A. JOHNSON, J.B., AND J.C. BAGLEY. 2011. Ecological drivers PUJADAS-BOTE, AND L. HERNÁNDEZ. 2009. El bosque of life-history divergence in poeciliid fishes. Pages tropical seco en riesgo: conflictos entre uso agro - 38–49 in J.P. Evans, A. Pilastro, and I. Schlupp, edi- pecuario, desarrollo turístico y provisión de servicios tors, Ecology and evolution of poeciliid fishes. Uni- ecosistémicos en la costa de Jalisco, México. Inter- versity of Chicago Press, Chicago, IL. ciencia 34:844–850. JOHNSON, J.B., AND M.C. BELK. 2001. Predation environ- ment predicts divergent life-history phenotypes CASWELL, H. 2000. Prospective and retrospective pertur- bation analyses: their roles in conservation biology. among populations of the livebearing fish Brachy - Ecology 81:619–627. rhaphis rhabdophora. Oecologia 126:142–149. ______. 2001. Matrix population models. Construction, JOHNSON, J.B., AND J.J. ZÚÑIGA-VEGA. 2009. Differential analysis, and interpretation. 2nd edition. Sinauer mortality drives life-history evolution and population dynamics in the fish . Associates, Sunderland, MA. Brachyrhaphis rhabdophora Ecology 90:2243–2252. CEBALLOS, G., AND A. GARCÍA. 1995. Conserving neotropi- JORDAN, F., K.J. BABBITT, AND C.C. MCIVOR. 1998. Sea- cal biodiversity: the role of dry forests in western sonal variation in habitat use by marsh fishes. Ecol- Mexico. Conservation Biology 9:1349–1353. ogy of Freshwater Fish 7:159–166. CHAPMAN, L.J., D.L. KRAMER, AND C.A. CHAPMAN. 1991. KILLGORE, K.J., J.J. HOOVER, S.G. GEORGE, B.R. LEWIS, Population dynamics of the fish Poecilia gillii (Poe- C.E. MURPHY, AND W. E . L ANCASTER. 2007. Distribu- ciliidae) in pools of an intermittent tropical stream. tion, relative abundance and movements of pallid Journal of Animal Ecology 60:441–442. sturgeon in the free-flowing Mississippi River. Jour- COGNATO, D.D., AND C.B. FIALHO. 2006. Reproductive nal of Applied 23:476–483. biology of a population of Gymnotus aff. carapo LEBRETON, J.D., K.P. BURNHAM, J. CLOBERT, AND D.R. (Teleostei: Gymnotidae) from southern Brazil. Neo - ANDERSON. 1992. Modeling survival and testing bio- tropical Ichthyology 4:339–348. logical hypotheses using marked animals: a unified CONTRERAS-BALDERAS, S., P. ALMADA-VILLELA, M.L. approach with case studies. Ecological Monographs LOZANO-VILANO, AND M.E. GARCÍA-RAMÍREZ. 2003. 62:67–118. Freshwater fish at risk or extinct in México. A check- MACÍAS-GARCIA, C. 1994. Social behavior and operational list and review. Reviews in Fish Biology and Fish- sex ratios in the viviparous fish Girardinichthys mul- eries 12:241–251. tiradiatus. Copeia 1994:919–925. DUNCAN, J.R., AND J.L. LOCKWOOD. 2001. Extinction in a MASEK, G.J., W.B. COHEN, D. LECKIE, M.A. WULDER, R. field of bullets: a search for causes in the decline of VARGAS, B. DE JONG, S. HEALEY, B. LAW, R. BIRDSEY, the world’s freshwater fishes. Biological Conserva- R.A. HOUGHTON, D. MILDREXLER, S. GOWARD, AND tion 102:97–105. W. B. S MITH. 2011. Recent rates of forest harvest and EVANS, J.P., A. PILASTRO, AND I. SCHLUPP, EDITORS. 2011. conversion in North America. Journal of Geophysical Ecology and evolution of poeciliid fishes. University Research Atmospheres 116:G00K03. of Chicago Press, Chicago, IL. MCKELLAR, A.E., M.M. TURCOTTE, AND A.P. HENDRY. GARCIA, A.M., J.P. VIEIRA, K.O. WINEMILLER, AND M.B. 2009. Environmental factors influencing adult sex RASEIRA. 2004. Reproductive cycle and spatiotempo- ratio in Trinidadian guppies. Oecologia 159:735–745. ral variation in abundance of the one-sided live- MEFFE, G.K., AND F. F. S NELSON JR., EDITORS. 1989. Ecol- bearer Jenynsia multidentata, in Patos Lagoon, Brazil. ogy and evolution of livebearing fishes (Poeciliidae). Hydrobiologia 515:39–48. Prentice-Hall, Englewood Cliffs, NJ. GHALAMBOR, C.K., D.N. REZNICK, AND J.A. WALKER. MIGUEL, A. 1991. Environmental pollution research in 2004. Constraints on adaptive evolution: the func- South America. Environmental Science & Technol- tional trade-off between reproduction and fast-start ogy 25:590–594. swimming performance in the Trinidadian guppy MILLER, R.R., W.L. MINCKLEY, AND S.M. NORRIS. 2005. (Poecilia reticulata). American Naturalist 164:38–50. Freshwater fishes of México. University of Chicago GIRONDOT, M., AND C. PIEAU. 1993. Effects of sexual dif- Press, Chicago, IL. ferences of age at maturity and survival on popula- OSPINA-ÁLVAREZ, N., AND F. P IFERRER. 2008. Tempera- tion sex ratio. Evolutionary Ecology 7:645–650. ture-dependent sex determination in fish revisited: 368 WESTERN NORTH AMERICAN NATURALIST [Volume 72

prevalence, a single sex ratio response pattern, and Gambusia affinis. Behavioral Ecology and Sociobiol- possible effects of climate change. PLoS ONE 3: ogy 61:1349–1358. e2837. SÚAREZ, Y.R., J.P. DA SILVA, L.P. VASCONCELOS, AND W. F. RAMOS-SANTIAGO, E., J.M. RAMÍREZ-GUTIÉRREZ, R. MEN- ANTONIALLI-JÚNIOR. 2009. Ecology of Phallotorynus DOZA-RODRÍGUEZ, AND M. TAPIA-GARCÍA. 2006. pankalos (Cyprinodontiformes: Poeciliidae) in a first- Reproducción, distribución y abundancia del pez order stream of the upper Paraná Basin. Neotropical Pseudupeneus grandisquamis (Perciformes: Mulli- Ichthyology 7:49–54. dae), en el Golfo de Tehuantepec, México. Revista SULLIVAN, J.A., AND R.J. SCHULTZ. 1986. Genetic and envi- de Biología Tropical 54:1103–1112. ronmental basis of variable sex ratios in laboratory REZNICK, D.N., M.J. BUTLER IV, AND F. H . R ODD. 2001. strains of . Evolution 40:152–158. Life-history evolution in guppies. VII. The compara- TURNER, C.L. 1937. Reproductive cycles and superfeta- tive ecology of high- and low-predation environ- tion in poeciliid fishes. Biological Bulletin 72:145–164. ments. American Naturalist 157:126–140. WHITE, G.C., AND K.P. BURNHAM. 1999. Program MARK: REZNICK, D.N., M.J. BUTLER IV, F.H. RODD, AND P. R OSS. survival estimation from populations of marked ani- 1996. Life-history evolution in guppies (Poecilia mals. Bird Study 46:S120–S139. reticulata) 6. Differential mortality as a mechanism WILLIAMS, B.K., J.D. NICHOLS, AND M.J. CONROY. 2002. for natural selection. Evolution 50:1651–1660. Analysis and management of animal populations. RICHARDSON, J.M.L., M.S. GUNZBURGER, AND J. TRAVIS. Academic Press, San Diego, CA. 2006. Variation in predation pressure as a mecha- WOOD, K.V., J.D. NICHOLS, H.F. PERCIVAL, AND J.E. nism underlying differences in numerical abundance HINES. 1998. Size-sex variation in survival rates and between populations of the poeciliid fish Heteran- abundance of pig frogs, Rana grylio, in northern dria formosa. Oecologia 147:596–605. Florida wetlands. Journal of Herpetology 32:527–535. SCHULTHEIS, C., A. BÖHNE, M. SCHARTL, J.N. VOLFF, AND YOUNG, J.A.M., J.R. MARENTETTE, C. GROSS, J.I. MCDON- D. GALIANA-ARNOUX. 2009. Sex determination diver- ALD, A. VERMA, S.E. MARSH-ROLLO, P.D.M. MAC- sity and sex chromosome evolution in poeciliid fish. DONALD, D.J.D. EARN, AND S. BALSHINE. 2010. Sexual Development 3:68–77. Demography and substrate affinity of the round SCOTT, L.E., AND J.B. JOHNSON. 2010. Does sympatry goby (Neogobius melanostomus) in Hamilton Har- environments predict life history and morphological bour. Journal of Great Lakes Research 36:115–122. diversification in the Mexican livebearing fish Poecili- ZULIAN, E., A. BISAZZA, AND G. MARIN. 1995. Variations in opsis baenschi? Biological Journal of the Linnean male body size in natural populations of Gambusia Society 100:608–618. holbrooki. Ethology Ecology & Evolution 7:1–10. SMITH, C.C. 2005. Sexual conflict and density depen- ZÚÑIGA-VEGA, J.J. 2011. Estimating potential reproductive dence in the western mosquitofish, Gambusia affinis costs in the survival of a xenosaurid lizard. Herpeto- (Poeciliidae). Master’s thesis, University of Ken- logical Journal 21:117–129. tucky, Lexington, KY. ______. 2007. Independent effects of male and female Received 29 February 2012 density on sexual harassment, female fitness, and Accepted 4 May 2012 male competition for mates in western mosquitofish