Litter size , effects of maternal body size , and date of birth in South American scorpions ( Arachnida : Scorpiones )

We present new data on litter size and date of birth (month) for 21 South American scorpions species. We provide data for one katoikogenic species, the liochelid Opisthacanthus cayaporum Vellard, 1932 (offspring = 3; birth month: Jan); and for several apoikogenic species, such as the bothriurids Bothriurus araguayae Vellard, 1934 (53; Sep), B. rochensis San Martín, 1965 (22-28; Jan, Aug); the buthids Ananteris balzanii Thorell, 1891 (10-34; Jan-Mar), Physoctonus debilis (Koch, 1840) (2; Sep), Rhopalurus amazonicus Lourenço, 1986 (19; Nov), R. lacrau Lourenço & Pinto-da-Rocha, 1997 (30; Dec), R. laticauda Thorell, 1876 (41; Nov), R. rochai Borelli, 1910 (11-47; Dec-Jan, Mar-Apr), Tityus bahiensis (Perty, 1833) (4-23; Oct-Mar), T. clathratus Koch, 1844 (8-18; Nov-Jan), T. costatus (Karsch, 1879) (21-25; Jan, Apr), T. kuryi Lourenço, 1997 (4-16; Mar), T. mattogrossensis Borelli, 1901(8-9; May), T. obscurus (Gervais, 1843) (16-31; Jan-Feb, May, Jul), T. serrulatus Lutz & Mello, 1922 (8-36; Dec, Feb-Apr), T. silvestris Pocock, 1897 (5-14; Dec-Jan, Apr), T. stigmurus (Thorell, 1876) (10-18; Nov, Jan, Mar), Tityus sp. 1 (T. clathratus group 7-12; Feb-Apr), Tityus sp. 2 (T. bahiensis group 2; Mar); and the chactid Brotheas sp. (8-21; Jan, Apr). We observed multiple broods: R. lacrau (offspring in the 2nd brood = 27), T. kuryi (6-16), T. obscurus (2-32), T. silvestris (8), T. stigmurus (4-9), T. bahiensis (offspring in the 2nd brood = 2-18; 3rd = 1), and T. costatus (2nd brood = 18; 3rd = 4). We found statistically significant positive correlation between female size and litter size for T. bahiensis and T. silvestris, and nonsignificant correlation for T. serrulatus.

Scorpions have many life history traits that are unusual for most terrestrial invertebrates (see reviews in POLIS & SISSOM 1990, LOURENÇO 2002a).Scorpions are viviparous contrasting with other Arachnida (FRANCKE 1982b, POLIS & SISSOM 1990, LOURENÇO 2002a, BROWN 2004).There are two types of embryonic development: apoikogenic and katoikogenic.Oocytes of apoikogenic species develop in ovarian follicules, which are attached directly to the ovariuterus, and are variable in size (small or large) and yolk content (rich or yolkless) (POLIS & SISSOM 1990, FARLEY 2001).Oocytes of katoikogenic species develop in specialized diverticula that branch from the female ovariuterus and are very small and yolkless, as such, the embryos are nourished via an oral feeding apparatus, a specialization of the diverticulum (POLIS & SISSOM 1990, FARLEY 2001).WARBURG & ROSENBERG (1996) described a third type of ovariuterus containing yolk-rich ova, in which oocytes apparently mature inside -rather than on the outer surface of -the ovariuterine tubes, forming a single row arranged in the shape of a bead necklace.
Gestation may be quite long, and many scorpion species are iteroparous, that is, capable of repeatedly giving birth after one or more matings (POLIS & SISSOM 1990).Some species of buthids are capable to produce multiple (1-5) broods after a single insemination by storing sperm (PIZA 1940, KOVOOR et al. 1987, LOURENÇO 2002a, ROUAUD et al. 2002).Storage of spermatozoa can greatly increase reproductive potential of the species (POLIS & SISSOM 1990, LOURENÇO 2002a).
Litter size in scorpions is poorly known; from over 1600 extant scorpion species, data are available for about 150 species.However, studies have shown that litter size is quite variable (FRANCKE 1981, POLIS & SISSOM 1990, BENTON 1991a, b, LOURENÇO et al. 1996, 2003, LOURENÇO & CUELLAR 1999, BROWN 2001, 2004, ROUAUD et al. 2002, LOURENÇO 2002a, 2007).WARBURG (2001) reported that it is difficult to assess the litter size in scorpions owing to the cannibalistic tendencies of the mother during parturition.Maternal cannibalism of stillborns and newborns could provide energy to the female, which nearly does not feed during the gestation period (PERETTI 1994).
ZOOLOGIA 26 (1): 43-53, March, 2009 Covariation of reproductive parameters such as egg size and clutch size with body size has been an area of intensive investigation in evolutionary ecology, since reproductive success is maximized, in part, by optimizing the number and size of young (SMITH & FRETWELL 1974, PARKER & BEGON 1986, LLOYD 1987).Empirical observation has demonstrated that larger females typically produce more and, somewhat less commonly, larger offspring (STEARNS 1992), and theoretical research predicts that clutch size and offspring size should exhibit a tradeoff (SMITH & FRETWELL 1974, STEARNS 1992).
Because they have more space available to store developing embryos or are better at obtaining, defending, storing, and allocating resources for reproduction, larger females often are predicted to produce more and larger offspring and to have greater reproductive investment (BROWN et al. 2003, BROWN 2004), and reproductive success (SKOW & JAKOB 2003).In most arthropod species, clutch size increases with female body (FOX & CZESAK 2000), suggesting a fitness advantage of larger females.
In several species of insects, fecundity -the total number of eggs or offspring an animal produces during each reproductive cycle (BEGON et al. 1990) or during the lifespan of a single female (YOSHIMURA 2003) -of larger individuals is greater than that of smaller ones (BERRIGAN 1991, PECKARSKY & COWAN 1991, HONEK 1993, YOSHIMURA 2003).Some studies indicate litter size is often positively correlated with female size in some species of spiders (MARSHALL & GITTLEMAN 1994, SIMPSON 1995, TANAKA 1995, PRENTER et al. 1999, PUNZO & HENDERSON 1999, BROWN et al. 2003, SKOW & JAKOB 2003), and in one species of solifuge (PUNZO 1998).
However, studies on reproductive allocation -how much of the available resources to dedicate to reproduction (reproductive investment) and how to allocate that resource fraction into offspring [per-offspring investment (BERNARDO 1996)] -in scorpions are rare (FRANCKE 1981, SMITH 1990, BENTON 1991a, b, FORMANOWICZ & SHAFFER 1993, BROWN & FORMANOWICZ 1995, 1996, LOURENÇO et al. 1996, 2003, LOURENÇO & CLOUDSLEY-THOMPSON 1999, BROWN 2001, 2004, LOURENÇO 2007).These studies indicate larger females generally produce larger litters and have a greater reproductive investment, although these trends do not hold for all species or even all populations of a single species (BROWN 2001(BROWN , 2004)).Moreover, scorpions seem to conform to CONGDON's (1989) optimal egg size theory for lizards, which postulates that variation in reproductive output results from variation in number of offspring rather than from differences in egg size (LOURENÇO et al. 1996, LOURENÇO & CLOUDSLEY-THOMPSON 1999).
Here, we present data on litter size and date of birth (month) of 21 species of South American scorpions within four families (Bothriuridae, Buthidae, Chactidae, and Liochelidae), in light of new information.In addition, we present new data on deferred fertilization for some Rhopalurus Thorell, 1876 and Tityus Koch, 1836 species.We also analyze statistically the relations between female size and litter size for three scorpion species: Tityus bahiensis (Perty, 1833), T. serrulatus Lutz &Mello, 1922, andT. silvestris Pocock, 1897 (Buthidae).

MATERIAL AND METHODS
Scorpions were collected under rocks or while active on the surface at night, using ultraviolet lights in different sites of Brazil, except for Bothriurus rochensis San Martín, 1965 (Bothriuridae) (Fig. 2) (N = 2), which were collected in Minas (Uruguai) (MZSP).Brazilian scorpions were: Ananteris balzanii Thorell, 1891 (Buthidae) (Fig. 3 The animals were kept in laboratory conditions at a mean temperature of 24ºC (20-28ºC), at the Departamento de Zoologia, Universidade de São Paulo.Each individual was housed in 15 x 15 x 6 cm plastic containers, with sand and humid cotton.They were fed crickets Gryllus assimilis (Fabricius, 1775) twice a month,  except when carrying offspring.Large scorpions were fed with nymph or adult crickets (about 20 mm total length) and the small ones with young crickets (less than 10 mm total length).
The containers were checked three times a week for occurrence of birth, presence of newborns, molted instars or dispersed offspring.Most females arrived pregnant from the field.Litter size (number of living and dead juveniles) were counted immediately after newborns were detected.To record litter size, newborns were removed, one by one, from female's dorsum and counted -we returned them to their mother immediately after data collection.The dead ones observed in the containers were counted as well and fixed in 70% ethanol.All females gave birth in laboratory conditions.Females' carapace length was measured using a dissecting microscope equipped with an optical micrometer.
We analyzed intraspecific relationships between female size and litter size for T. bahiensis, T. serrulatus, and T. silvestris the three species for which we collected sufficient data (n = 35, n = 10, and n = 9, respectively), using least-square regression.For the analyses, alpha was set at 0.05.Data analyses were carried out using SPSS for Windows version 14.0.
We observed multiple broods in seven buthid species (Tab.II) probably because of the capacity of some species of buthids to store sperm (PIZA 1940, KOVOOR et al. 1987, LOURENÇO 2002a, ROUAUD et al. 2002).We provide the first record of this phe- nomenon for the genus Rhopalurus (R. lacrau), and for some Tityus species (T.costatus, T. kuryi, and T. silvestris).

Litter size
We observed a small brood size (3) for the katoikogenic O. cayaporum (Liochelidae) (Fig. 1), while POLIS & SISSOM (1990), LOURENÇO (1991LOURENÇO ( , 2002a) ) and LOURENÇO et al. (2003) had reported 15-25.In most arthropods, progeny size decreases with maternal age (FOX & CZESAK 2000).Decrease of litter size can be observed for females of any scorpion species which arrive to the  end of their life cycle, when the number of follicles starts to decline (LOURENÇO 1979, WARBURG et al. 1995, WARBURG & ELIAS 1998a, b).However, we are not sure whether this is the case we observed since life span of O. cayaporum is, in average, seven years (LOURENÇO 2002a, LOURENÇO et al. 2003).As the observed female lived as an adult during four years in our laboratory and gave birth in the first year, we can estimate she was in the middle of her life cycle when she gave birth.Furthermore, katoikogenic embryos may have some greater control over the energy invested in their development than do apoikogenic embryos (POLIS & SISSOM 1990), so that the former species might be predicted to produce relatively small litters (BROWN 2001).For Bothriuridae, PERETTI (1997) observed an average number of 32 ± 1.40 offspring per parturition and a maximum number between 46-55 that occurs in 18% of the cases.Moreover, most of the bothriurid females had similar litter size (VARELA 1961, MAURY 1969, 1978, MATTHIESEN 1971b, ACOSTA 1983, POLIS & SISSOM 1990, PERETTI 1997).In our observations, B. araguayae (53 offspring) had larger litter size than reported by MATTHIESEN (1971b -28 offspring) and most bothriurid species, presenting one of the highest values of litter size for the family, whereas B. rochensis (22-28) (Fig. 2) produced slightly smaller litters than most of the bothriurid species.
LOURENÇO & CUELLAR (1999) observed an average number in brood of 16 for A. coineaui.Most of the small buthid species show strongly reduced average litter size, in most cases with numbers less than 10 (ROUAUD et al. 2002, LORENÇO 2007).Thus, A. balzanii (10-34 young) (Fig. 3) produced in average litter size larger than small species of buthids, such as A. coineaui, and most species of Tityus (15-25, see ROUAUD et al. 2002).MATTHIESEN (1968MATTHIESEN ( , 1971b)), POLIS & SISSOM (1990) and LOURENÇO (2007) reported litter size of 28-49 for R. rochai and SARMENTO et al. (2008) reported an average of 35.80 ± 12.11 (23-55) offspring.Although observed an average litter size (33 ± 11.90) (Fig. 4) similar to that presented by SARMENTO et al. (2008), our data (11-47) provide smaller numbers than those published by these authors.When we detected the presence of newborns from this female in particular, four of her 11 offspring were dead.Hence, there could be more stillborns that we did not record due to maternal cannibalism (a behavior unobserved by us) before checking the container.SARMENTO et al. (2008) recorded cannibalism by the female in four out of the five observations.This is the first record on litter size for T. kuryi (4-16).One particular female gave birth to four offspring, which is much smaller than litter size of most species of Tityus (15-25, see ROUAUD et al. 2002) and the average number for the order (26, see POLIS & SISSOM 1990).AGUIAR et al. 2008 reported litter size of 10.40 ± 2.90 (6-16) for T. stigmurus.These authors also observed that litter size and offspring mass were unrelated to female size, although there was significant positive correlation between total litter mass and litter size.Thus, female reproductive resources in T. stigmurus appeared to be applied to the production of more but not heavier offspring.Our data [13 ± 4.36 (10-18)] provide a slightly larger brood sizes than those published by these authors and litter size similar to most species of Tityus (15-25, see ROUAUD et al. 2002).(2001), LOURENÇO (2002a, 2007), ROUAUD et al. (2002), and LOURENÇO et al. (2003), litter size of Tityus species we reported in our laboratory are similar to those published.Closest similarities were found between T. clathratus (8-18), T. mattogrossensis (8-9), T. silvestris (5-14), Tityus sp. 1 (7-12) (Fig. 7), and other species of T. clathratus group (12-13).
Within Chactidae, this is the first record for litter size and date of birth for the genus Brotheas (8-21 offspring) (Fig. 8).However, LOURENÇO et al. (2003) had reported litter size of 14-15 for Chactas lepturus (total length: 50-55 mm) and 15-16 for C. reticulatus (length: 50-60 mm).Thus, Brotheas sp.produced larger litter size than both species of the genus Chactas in spite of similar body size.
Furthermore, laboratory-reared individuals often suffer extremely high mortality (90-100%) (POLIS & SISSOM 1990), due to stillborns, failure of newborns to detach from the birth membranes or to ascend to female's dorsum, maternal cannibalism, or failure to feed or molt successfully (PIZA 1940, STAHNKE 1966, FRANCKE 1976, 1979, 1982a, POLIS & SISSOM 1990).Maybe the smaller values of litter size observed in the present study, compared to data from literature, is due to a decline of the number of follicles (in the end of the individual life cycle) (LOURENÇO 1979, WARBURG et al. 1995, WARBURG & ELIAS 1998a, b), oosorption [known to occur in scorpions (MAURY 1969, WARBURG et al. 1995, WARBURG 2001)], or is the result of maternal cannibalism of stillborns or newborns that failed to ascend to the female's dorsum or -for apoikogenic species -failed to detach from the birth membranes [membranes of katoikogenic species disappear before parturition (FRANCKE 1982b)].In our laboratory, we have already observed occurrence of abortion (R. agamemnon), stillborns (B.araguayae, B. rochensis, R. rochai, T. bahiensis) and maternal cannibalism of newborns (T.bahiensis), probably due to stress conditions (FOX 1975).MAURY (1969) and PERETTI (1994) noted that maternal cannibalism increased under stress conditions, especially during birth and newborns.Moreover, POLIS & FARLEY (1980) and WARBURG (2001) reported that it is always possible (probable) that some juveniles have disappeared (by oosorption and/or maternal cannibalism) during the process of parturition, thus the number of juveniles eventually born is considerably smaller.In addition, laboratory conditions might also have been more stressful than in the field, leading to decreased investment in reproduction.

Date of birth
We provide evidence that the great majority of births occurred during South American warm season, throughout the months of September to April; with most births occurring during the summer (December to March, Tab.I).Opportunistic and synantropic species, such as T. bahiensis and T. serrulatus are able to give birth throughout the year (POLIS 1990, LOURENÇO & CUELLAR 1995, EICKSTEDT et al. 1996, LOURENÇO 2002a, LOURENÇO et al. 2003).Hence, the ability to give birth throughout the year can greatly increase their population in the cities and the number of accidents with humans (LOURENÇO 2002a).

Deferred fertilization
We observed multiple broods in seven buthid species (Tab.II), including the first record of this phenomenon for the genus Rhopalurus (R. lacrau), and for some Tityus species (T.costatus, T. kuryi, and T. silvestris).We noticed a decrease in litter size through multiples broods, except in T. silvestris, and one case in T. bahiensis and T. kuryi.This could be explained by a reduction on viability of stored spermatozoa, or by the ageing of the female.PIZA (1940), BÜCHERL (1956), and MATTHIESEN (1961, 1968) reported the occurrence of storage of spermatozoa in T. bahiensis, being 7-20 the number of offspring per female (MATTHIESEN 1961) and the capability of giving birth to three broods after mating (MATTHIESEN 1968).We observed seven out of eight females giving birth to two broods and one of them gave birth to a third brood, being 1-21 the number of offspring per female.Our data corroborate MATTHIESEN's number of broods and shows a lower number of offspring per female.MATTHIESEN (1971b) observed that three females of T. obscurus produced multiple (2-3) broods, but he did not mention the number of young born from each parturition.In our observations, two females gave birth to two broods: one female gave birth to 21 and two offspring respectively in the first and second broods, and the other gave birth to 31 and 32 offspring respectively.MATTHIESEN (1971b) observed one female of T. stigmurus giving birth to three broods in laboratory, but he did not mention litter size.We observed two females giving birth to two broods, with 4-18 offspring per female.
Although BROWN (2001) found little evidence for the covariation between female size and litter size at the intraspecific level in scorpions, with the increase of research in this area, more evidence has been found to support such prediction (see BROWN 2004).Moreover, variation in female size explained 19.4% of the variation in litter size for T. bahiensis (R 2 = 0.194) and 73.3% for T. silvestris (R 2 = 0.733).Thus, larger females of T. bahiensis and T. silvestris produced more offspring (had larger litter) and invested more into reproduction.In addition, larger females have more room to store developing young, that is, if all females obtain sufficient resources for reproduction, then clutch size may be affected by available space within the female for developing offspring (see BRADLEY 1984, SINERVO & LICHT 1991).However, correlation between female size and litter size for T. serrulatus was nonsignificant (p = 0.373; F = 0.889), although larger females tended to have larger litters (Fig. 10).This trend is common for most scorpion species (BROWN 2001, 2004), but WARBURG (2001) reported that larger females do not always produce more embryos; it may be related to the physiological condition of the female, instead of female size.
A second explanation for the lack of correlation between female size and litter size for T. serrulatus involves the lower variance in sizes presented by the females in our samples.Alternatively, our sample size is just too small to reveal effects of female size in litter size and further studies should be conducted, although BROWN (2001BROWN ( , 2004) ) reported that, among scorpions, sample size is uncorrelated with the magnitude of the correlation coefficient between female size and litter size.
Intraspecific and individual variation in reproduction can be due to other factors besides female size such as foraging success, temperature, humidity, photoperiod, age, and physiological state of the female (see FISCHER & VASCONCELLOS-NETO 2005).Thus, differences in resource availability and acquisition among females -which can occur if better quality females obtain more or better quality prey -may also explain the trends observed (BROWN & FORMANOWICZ 1995, 1996).Besides, LOURENÇO et al. (1996) observed that a larger litter could be due to greater food availability and POLIS & SISSOM (1990) reported that low food levels might lead females to invest fewer resources in reproduction, or to reabsorb some embryos (POLIS & SISSOM 1990).On the other hand, if all females obtain sufficient resources for reproduction, then clutch size may be affected by available space within the female for developing offspring (see BRADLEY 1984, SINERVO & LICHT 1991).Even though females of T. serrulatus were collected in different sites and date, most (50%) were from two close localities (Bahia State, Ceraíma and Guanambi), and this bias could have influenced our results.Moreover, T. serrulatus is partenogenetic and is capable of producing multiple broods during a single year.If litter size changes during the breeding season and the life cycle, then the values used for this species may not accurately reflect the true mean litter size (see BROWN 2001).Another factor that may also be involved is that laboratory conditions are more stressful than in the field, and might also have influenced these trends by leading to distinct (decreased) investment in reproduction.
Finally, spatial or temporal variation in the intraspecific relationship between female size and litter size suggests that patterns of resource allocation (to reproduction) are not fixed in a species, but may vary due to genetic differentiation (local adaptation) among populations or due to environmental stochasticity among years (BROWN 2001).More information on life histories in other scorpions needs to be gathered to determine the general trend of the correlation between litter size and female size within the order.

.14 Tityus sp.2 (bahiensis group) 1 2 Mar 6.07 Chactidae Brotheas sp. 3 8-21 Jan and Apr 7.89 ± 0.73
(LOURENÇO 2002b)als present the dentate margins of pedipalp-tibia fingers composed of 16 oblique rows of granules.According to the key to the species of the group T. clathratus(LOURENÇO 2002b), these correspond to T. mattogrossensis, although T. pusillus Pocock, 1893 is the only species of the group known so far from the state of Pernambuco.

Table II .
Deferred fertilization in South American Buthidae species.New data are in bold.

Table III .
Parthenogenesis in Tityus serrulatus.Four females gave birth to two broods.