Study site

The toads were collected from 5 sites along the Honken River in Deh-Ken (24° 10′N, 120° 43′E), Taichung City, central Taiwan. Data for temperature and rainfall for 1990–1991 in central Taiwan are summarized in Fig. 1 (Data are from the Shuijaitou Weather Station, Central Weather Bureau, ROC). According the wild observations, such as mating call, amplexus behavior, tadpole emergent time, and collection number of toads, four periods in an annual cycle, February to April, May to June, July to September, and October to January, were determined in the present study.

Fig. 1.

Monthly changes in rainfall and mean air temperature in Taichung, Taiwan.

Collection of blood, tissue and organs

Toads were collected at night from the 5 sites on the 5th, 15th, and 25th days of each month from March 1990 through March 1991. The numbers of toads collected each month are shown in Table 1. In the torpid period, it was difficult to find toads, especially in July. The collections of blood, tissue and organs were done in the laboratory on the next day following capture of the toads. All toads were weighed and measured before dissection; the fresh weights of testes, fat bodies, and liver were recorded. To minimize the body weight differences of individual toads, weights of organs were also expressed as an organosomatic index as (gram organ weight/gram body weight) × 100%.

Table 1

Changes in the mean weights of the body, testis, liver, and fat body during an annual reproductive cycle of male Bufo melanostictus

Blood was collected in heparinized microtubes from the conus arteriosus after the toads were anesthesized with diethyl ether. The blood sampling was usually completed within 10 min for each individual. The blood was centrifuged at 10³ × g for 10 min, and plasma (0.5 to 1.2 ml) was collected and then stored at −20°C unti the assay of androgen.

Histological examination

The right testis of each animal was fixed in Bouin's solution, embedded in paraffin, serially sectioned at 6 μm, and stained with hematoxylin and eosin. Spermatogenetic activity was assessed as described by Saidapur (1983): Stage 0, primary spermatogonia; stage I, secondary spermatogonia; stage II, primary spermatocytes; stage III, secondary spermatocytes; stage IV, presence of sperm bundles in the seminiferous tubule; and stage V, appearence of mature sperm in the seminiferous tubule lumen. The intensity of spermatogenetic activity is expressed by the number of sperm bundles in the seminiferous tubules. Sperm bundles were coounted from 20 random sections of each testis. The number of sperm in the seminiferous tubule was calculated as the area of free sperm in the seminiferous tubule lumen divided by the total area of seminiferous tubule lumen; values are expressed as abundant (> 50%), median (30%–50%) and few (< 30%).

Radioimmunoassay of androgen levels in plasma

Plasma androgen was measured by a radioimmunoassay described previously by Kao et al. (1993). Androgens in the samples were extracted with diethyl ether; the average recovery of the added testosterone was 75.3 ± 4.3%. The cross-reactivites of the antiserum with steroids were tested for testosterone (100%), dihydrotestosterone (74%), androstenedione (1.23%), and androstenediol (0.59%). Thus, the data are expressed as “androgen” which mainly represents testosterone and dihydrotestosterone present in the plasma. Intra- and inter-assay variabilities had been previously established at 6.7% and 12.9%, respectively.

Statistical analysis

Analysis of variance (ANOVA) and Duncan's multiple range tests were used to examine differences among the numerical mean values. Pearson's correlations analyses were performed for all variables (SAS, 1994). Regression analysis was performed between SB, FBSI, LSI, as well as TW (for the explanations of abbreviations, see Table 4), and environmental variables. In the analyses, SB, FBSI, LSI, AL, and TW were treated as dependent variables, whereas the climatic factors of monthly air temperature and rainfall were treated as independent variables. A probability value of 0.05 or less was considered to indicate significance.

Observations of reproductive cycle, and climatic factors

The mating call and amplexus were observed in February and tadpoles appeared at the end of March. The breeding season continued through May; the numbers of adults and tadpoles decreased after the end of May and the tadpoles disappeared in June. In October–December, when temperatures decreased to 17°C, it was difficult to find and collect toads. As shown in Fig. 1, the highest mean temperatures were in July 1990, at 28°C, with a decrease to the lowest at 17°C in January 1991. The period between April and August 1990 was the rainy season.

Histological observations of testes

The histological examinations of B. melanostictus testes collected monthly from March 1990 through March 1991 revealed that the spermatogenetic activities in the months of January–March, June, August, and October–December were at stage V; while in April, May, July, and September the activities were at both stage IV and stage V (Table 2).

Table 2

Most advanced stage of spermatogenesis found in testis of Bufo melanostictus in different months of the year

The number of sperm bundles increased from June 1990, then declined to the lowest in December, and peaked in March 1991 (Table 3). In the 4 annual periods (February to April, May to June, July to September, and October to January), the mean numbers of sperm bundles were 6.3, 5.2, 5.5, and 3.6, respectively. No significant difference was found among these 4 periods (F_{3, 9} = 1.01, p = 0.43), suggesting that spermatogenic intensity does not vary significantly throughout the year. The numbers of sperm bundles were significantly correlated with LSI (r = 0.56, p < 0.05) and plasma androgen (r = 0.82, p < 0.05) (Table 4).

Table 3

Monthly change of spermatogenetic intensity: number of sperm bundles (SB, mean ± SE) and number of sperm in seminiferous tubules (SN) of Bufo melanostictus

Table 4

Correlation coefficients (r) of sperm bundles (SB), fat body somatic index (FBSI), liver somatic index (LSI), level of plasma androgen (AL), testis weight (TW), temperature (TEM), and rainfall (RF) of male Bufo melanostictus during an annual reproductive cycle

Annual changes in TW, LSI, and FBSI

As indicated in Table 1, the mean testis weight (TW) showed significant variations between months (F_{1, 9} = 5.92, p < 0.05) and had significant correlations with both sperm bundles (r = 0.65, p < 0.05) and plasma androgen (r = 0.78, p < 0.01) (Table 4). The mean liver somatic indices (LSIs) were 2.4%, 3.1%, 2.8%, and 3.1% (Fig. 2A) for the periods February–April, May–June, July–September, and October–January, respectively, and showed no significant difference (F_{3, 9} = 0.99, p = 0.44). The mean fat body somatic indices (FBSIs) were 0.3%, 0.5%, 3.1%, and 5.1% (Fig. 2B) for the periods October–January, February–April, May–June, and July–September, respectively, and also showed no significant variations among the 4 periods (F_{3, 9} = 1.18, p = 0.37).

Fig. 2.

Monthly changes of the Ivier somatic index (A), and fat body somatic index (B). The data are expressed as mean ± SE, and the numbers above the dots indicate monthly sample sizes.

Annual cycle of plasma androgen levels

Plasma androgen levels peaked in March (mean value=12.0 ng/ml), and then decreased to the lowest in December (0.5 ng/ml) (Fig. 3). A significant periodic trend in plasma androgen levels was recorded (F_{3, 9} = 3.96, p < 0.05).

Fig. 3.

Monthly changes of plasma androgen levels during an annual reproductive cycle of Bufo melanostictus.

The correlations among sperm bundles (SB), LSI, FBSI, levels of plasma androgen (AL), TW, temperature, and rainfall are summarized in Table 4. As indicated, AL was positively correlated with SB and TW.

Reproductive cycle

The reproductive cycle of B. melanostictus may be divided into 4 periods: breeding, post-breeding, reproductive energy preservation and torpid periods. The numbers of sperm bundles in the seminiferous tubules, the numbers of free sperm, and plasma androgen levels in February–April were all higher than during the other 3 periods. Field observations showed that the breeding behaviors occurred in February–April. Thus, February–April is considered to be the breeding period. In May–June, the numbers of sperm bundles, free sperm, and adult toads were lower, and this is considered to be the post–breeding period. In July-September, the numbers of sperm bundles and free sperm were low; the FBSI value in this period was indeed higher than in the other three periods. July–September was thus considered the reproductive energy preservation period. In October–January, the numbers of sperm bundles and free sperm were lowest; at the same time, adult toads were difficult to find in the field. This period was thus considered the torpid period.

Testicular activity

Amphibians inhabiting tropical areas, where climatic conditions do not show appreciable fluctuations, have often developed a continuous type of spermatogenic cycle; whereas other species in temperature regions have developed a discontinuous or potentially continuous type of spermatogenic cycle (Basu and Nandi, 1965; Lofts, 1974). In species with a continuous type of spermatogenic cycle, the spermatozoa are generally produced throughout the year and the testes always contain spermatic cell nests as well as a complete spectrum of spermatogenic stages (Sun, 1979; Saidapur and Kanamadi, 1982; Saidapur, 1983). In contrast, the spermatogenic activity of species living in temperature regions in restricted to late spring and summer; such discontinuous production of spermatozoa has been attributed, at least in part, to a seasonal lowering of environmental temperatures and an accompanying decline of gonadtropin output (Lofts, 1974). Mondal and Basu (1960), Lofts (1974), and Saidapur (1983) proposed that the spermatogenic cycle of species which have free sperm in the seminiferous tubules belongs to the continuous type. In other words, the classification of spermatogenic cycles of anuran species belonging to continuous or discontinuous types depends on the presence or absence of free sperm in the seminiferous tubules. As observed in the present study, B. melanostictus had free sperm in the seminiferous tubules the entire year. According to the categorizations by Lofts (1974) and Saidapur (1983), the spermatogenic cycle of B. melanostictus thus belongs to the continuous type. However, a comparison of the numbers of free sperm showed that they were significantly higher in the breeding period than during any of the other 3 periods (Table 3). Such observations reveal that although B. melanostictus belongs to the continuous type of spermatogenic cycle, it has a distinct breeding period. Accordingly, it is inappropriate to determine the nature of the reproductive cycle of toads by depending solely on the presence of free sperm in the seminiferous tubules. We thus propose that the continuous type of spermatogenic cycle of anurans may be divided into the constantly continuous type (e.g., R. cyanophlyctis; Saidapur and Kanamadi, 1982), and the fluctuating continuous type (e.g., B. melanostictus, present study).

Changes in the weights of the testes have also been used as a basis to determine the reproductive cycle of amphibians (Lofts et al., 1972; Rastogi et al., 1986). The results of the present study reveal that the annual variations of testicular weight of B. melanostictus show a significant difference between months, and that testicular weight was significantly correlated with the number of sperm bundles and plasma androgen levels (Table 3). Thus, seasonal variations in testicular activity are generally reflected in gonadal weight. Such a pattern is generally similar to that observed in B. bankorensis, a species exhibiting a potentially continuous type of spermatogenic cycle (Huang et al., 1996), and to that of species with a discontinuous-type testicular cycle, such as R. temporaria, in which testicular weight is an appropriate index of testicular function (Lofts et al., 1972).

Androgen patterns and correlation with reproductive cycle

The patterns of seasonal plasma androgen levels of B. melanostictus observed in the present study are very similar to those previously reported Dicroglossus occipitalis (Kuhn et al., 1987), B. japonicus (Itoh and Ishii, 1990a), Rana rugulosa (Kao et al., 1993) and B. bankorensis (Huang et al., 1996); all of these anurans exhibited a distinct circulating androgen peak during their breeding period. It is usually accepted that androgen is related to the initiation of breeding activity and the expression of sexual behavior (Duellman and Trueb, 1986). Although the exact functions of androgen in B. melanostictus are unknown at present, the finding that plasma androgen levels are maximal in the breeding period suggtests that the mating behavior is also probably androgen-dependent in this species. Siboulet (1981) also demonstrated that circulating androgen levels in male B. mauritanicus are higher during periods of amplexus.

In some of the anurans investigated, a notable decrease in plasma androgen levels and an increase in seprmatogenic activity were observed after the breeding period [e.g., B. mauritanicus (Siboulet, 1981) and Pachymedusa dacnicolor (Rastogi et al., 1986)]. A possible androgen negative feedback mechanism at the hypothalamus-hypophysis axis level has been proposed (Rastogi et al., 1976; Delgado et al., 1989; Itoh and Ishii, 1990a). It is generally believed that in early phases of the spermatogenic cycle of anurans, the mitotic proliferation of primary and secondary spermatogonia is gonadtropin-independent (Lofts, 1974), and that the meitotic proliferation of secondary spermatocytes, formation of sperm bundles, and presence of free sperm are androgen-dependent (Saidapur, 1983). The present study of B. melanostictus revealed that the higher levels of androgen during the breeding period from February to April were observed coincident with the increase of sperm bundles and free sperm in the seminiferous tubules. Such observations are consistent with those reported for other anurans (Rastogi et al., 1976; Kobayashi and Iwasawa, 1986; Saidapur, 1983; Kao et al., 1993; Huang et al., 1996).

Relations of changes of fat body and liver weights to the reproductive cycle

Lipids are the most efficient storage form of the 3 potential energy compounds (lipids, glucose, and proteins) in anurans (Long, 1987). The patterns of annual lipid variation have been observed in many anurans; body lipid reduction may reflect its metabolic use during dormancy, for vitellogenesis, and possibly for steroid production during the reproductive period (Chieffi et al., 1975; Rastogi et al., 1978; Morton, 1981; Pramoda and Saidapur, 1984; Kanamadi and Saidapur, 1988). As observed in the present study, no significant correlations were observed between FBSI and various reproductive parameters (Table 4); the role of fat bodies during the annual reproductive cycle of B. melanostictus is still nuclear and requires further investigation. Liver glycogen also serves as a metabolic substrate in anurans (Brenner, 1969). The negative correlations between LSI and sperm bundles observed in the present study suggest that the energy accumulation in the liver may be used in the spermiogenesis activity of B. melanostictus.

Association of environmental factors with the reproductive cycle

Three major environmental factors have been implicated in the regulation of the amphibian breeding cycle: rainfall, photoperiod, and temperature (Lofts, 1974). Rainfall has been shown to influence the breeding behavior in various anurans (Lofts, 1974). It was demonstrated that photoperiod potentially modifies testicular activity in R. esculenta (Rastogi et al., 1976). Temperature, in contrast, has been shown to play an important role in regulating plasma androgen levels and in synchronizing the different phases of the seasonal testicular cycle in several anurans (Rastogi et al., 1976, 1978; Iela et al., 1980; Pancak and Taylor, 1983; Pierantoni et al., 1985). It was further affirmed that temperature was the most important factor and that photoperiod plyaed only a permissive role in the mediation of temperature influence on the gametogenic and endocrine activity of R. esculenta from the temperate region (Rastogi et al., 1978). Our present study of the subtropical anuran species, B. melanostictus, revealed that no apparent correlations existed between temperature and testicular activities or between rainfall and testicular activities. The relation of photoperiod to testicular activities of the toad was not investigated in this study.