The three studied species showed similarities and differences in the reported population parameters. In terms of population size, H. atalanta was dominant, followed by C. eous and M. affinis. This parameter can be directly related to its host plant use, with H. atalanta using the predominant grass species in the study area, and C. eous and M. affinis using host plants that are less abundant and with restricted distributions in the study area. Concerning the variation in numbers through the year, however, H. atalanta and C. eous exhibited a similar pattern, with somewhat stable populations during the wet season and beginning of the dry season, with a reduction in numbers towards the dry season, while M. affinis showed a single population peak followed by very low numbers during most of the dry season.
Apparently, population decreases are not related to lawn mowing, as reported for Eurema elathea (Pieridae) in the same study area (Vanini et al. 1999), except for the five successive mowing events in early June 1999, which may have contributed to a pronounced population decrease in all three species (see Fig. 2). Given that the three studied species breed on grasses that are mostly growing in the grass lawns (see above), local mowing likely destroyed their immature stages and possible contributed to adults moving to another sub-areas, and this result is especially critical if mowing occurs in periods of low population numbers or near the periods of population decline (Vanini et al. 1999; Tourinho and Freitas 2009).
Male-biased sex ratios were reported for H. atalanta and C. eous, but not for M. affinis. Male-biased sex ratios are usually reported in population studies of tropical and temperate butterflies, including several species of satyrines (see Seixas et al. 2017; Rosa et al. 2020 and references therein). Usually, behavioral differences between sexes are argued to be the reason for this pattern, with males and females flying in different parts of the habitat (Ehrlich and Gilbert 1973; Mallet and Jackson 1980; Freitas 1993). In the present study, however, virtually the entire habitat was covered during MRR sampling in the three sub-areas. Females could have larger home ranges, moving to nearby grass lawns searching for better ovipositing sites (a hypothesis supported by the lower recapture rates of females in both species) (Wiklund 1977), and therefore may have a broader home range than males. However, for every female that flies out of an area, one should expect another to fly in from a nearby area, resulting in no changes in sex ratios. In contrast, M. affinis presented a sex ratio of 1:1 and the recapture rates were equal for both sexes. This could be explained by the high philopatry of this species, which was basically restricted to a sector of sub-area 3 where its host plant M. maximus thrives, allowing the capture of most individuals in the population (evidenced by the high recapture rate reported in this species compared with H. atalanta and C. eous). Although the low sample size could affect the detection of a male biased sex ratio in this species, sex ratio of M. affinis in laboratory rearing was also 1:1 (AVLF unpublished data).
Age structures of all three species are similar in respect that there are no marked events of recruitment of new individuals in any specific time of the year, which is expected for species that have continuous recruitment throughout the year and overlapping generations (Ramos & Freitas 1999). The average permanence (an indirect measure of lifespan) was similar for H. atalanta and C. eous (about one week, with maximums of nearly a month), and slightly higher for M. affinis (about 11 days for males, but also with a maximum of nearly a month). However, the life expectancy was lower in H. atalanta when compared with the other two species, which can be visually verified in the survival curves: although similar, the survival curve for H. atalanta shows a pronounced decline in adults above the category of 16–20 days of permanence when compared with C. eous and M. affinis.
The mean adult lifespans of about one week reported here are noticeably short in comparison with other tropical butterflies, including Heliconius, some Ithomiini, Papilionidae and even other fruit-feeding nymphalids, including some euptychiines, whose adults present mean lifespans considerably longer, from two weeks to one month or more (Ramos and Freitas 1999; Andrade and Freitas 2005; Seixas et al. 2017; Pedrotti et al. 2019; Lourenço et al. 2022). Although several factors could underly for butterfly longevity, adult feeding has been considered one of the main factors influencing this parameter, especially in fruit-feeding butterflies (Bauerfeind and Fisher 2005; Molleman et al. 2007). Although fruit trees, such as guava (Psidium guajava L.; Myrtaceae) and black mulberry (Morus nigra L.; Moraceae), are present in the campus, these are not abundant and are located near to buildings and fences. Thus, the scarcity of food resources (especially fleshy fruits, see above) throughout most of the year in the study area could explain the short lifespans reported for the three species here studied. However, there are no comparative studies of these same species in forested habitats for comparison, and the low number of population studies with tropical butterflies prevents the identification of general patterns for different groups.
The abundance of each species by sub-area are not related to total area, and only C. eous presented higher abundance in area 1 (the largest sub-area in the present study). Thus, abundances are likely explained by biological factors such as microhabitat and presence of host plants. For example, the M. affinis is mostly restricted to the small, fenced sector inside of area 3 (Fig. 1D), where its host plant M. maximus persist due to the very low management (mechanized mowing was never observed inside the fence). Considering densities, the higher valor for C. eous was reported in area 2, where the shaded conditions imposed by the buildings allowed the persistence of one of its host plants, E. indica, a grass species that is especially common in the contacts of paved areas with the grass lawns, and area 3, where both host plants persist due to the absence of mechanized mowing and shaded conditions. In addition, C. eous is a territorial species (Peixoto and Benson 2008, 2009), and the spatial distribution of males could be partly explained by the distribution of potential defendable territories. Previous studies showed that males of this species establish territories in sunny clearings at forest edges (Peixoto and Benson 2008, 2009), a condition that is absent in the open sunny lawns (explaining the low densities of C. eous in area 1), but present in the area 2 and especially in the fenced sector of area 3. In the case of H. atalanta, however, the host plant P. notatum is the main grass in all studied sub-areas, and other factors should explain the higher densities in area 3 and area 2, respectively. In this case, high densities could be related to local environmental conditions such as microclimates, presence of adult resources (such as scarce fermenting fruits and other occasional decaying organic material), or more suitable larval host plants.
In the present study, two species, C. eous and M. affinis, present female-biased sexual size dimorphism, while this bias was not reported for H. atalanta. In fact, insects in general present a female-biased sexual dimorphism in size (Stillwell et al. 2010), a feature usually linked with greater fecundity in bigger females (Allen et al. 2011 and references therein). This pattern was also that most frequently reported for butterflies (Allen et al. 2011), with the remarkable exception of species in the genus Heliconius, whose males do not differ in size or are larger than females (Ramos and Freitas 1999; Andrade and Freitas 2005; Seixas et al. 2017; but see Hernandez and Benson 1998). Another exception includes territorial butterflies, whose larger males have an advantage in confrontations, resulting in males usually being larger than females (Allen et al. 2011). However, although C. eous is a territorial species, body mass and not wing length was not reported as a reliable predictor of success in territorial contests (Peixoto and Benson 2008), which could explain the larger females in this species.