Introduction

The origin of eusociality, in which a colony includes nonreproductive individuals (usually referred to as workers), has attracted the interest of both evolutionary biologists (Futuyma and Kirkpatrick 2017) and sociobiologists (Wilson 1975). A presumably fruitful approach for understanding the origin of workers is to elucidate the mechanisms underlying caste-fate determination in primitively eusocial species (Oster and Wilson 1978; Hunt 1991; Smith et al. 2008) in which there are no critical morphological differences between the queen and workers (Michener 1964; Wilson 1971; Jeanne 2003) because primitive eusociality appears to represent an early stage of eusociality (Wilson 1971; Wheeler 1986; Hunt 2012; Boomsma and Gawne 2018; Piekarski et al. 2018). All individuals of these species have the capability of reproducing, and levels of sociality are less developed than in advanced eusocial organisms in terms of colony size and conflict and cooperation among the queen and workers. In particular, primitively eusocial groups in Polistinae are suitable models for exploring the evolution of eusociality because their sociality originated in the shared ancestor of the subfamilies Polistinae and Vespinae (Piekarski et al. 2018), which include advanced eusocial species as well as primitive eusocial species. Thus, studying primitive eusociality in paper wasps, such as Polistes, can provide insight into some of the possible traits involved in caste determination (Jandt and Toth 2015).

It has often been suggested that Vespidae evolved eusociality through the following steps (Pardi 1948; West-Eberhard 1978; Carpenter 1991; Gadagkar 1991; Field and Foster 1999; Noll et al. 2004; Hunt 2012; Piekarski et al. 2018): (1) solitary habits, (2) casteless nest sharing, (3) eusociality with behavioral castes that do not include preimaginal caste-related biasing (PCB), (4) eusociality with PCB, and (5) eusociality with morphologically differentiated castes. The development of sociality might originate from the first brood comprising nondiapausing adults in temperate groups (diapause ground-plan hypothesis; Hunt and Amdam 2005; Hunt 2006, 2007, 2012). However, this assumption is challenged by the following two facts, as suggested by Kelstrup et al. (2017): (1) the phylogenic and biogeographical analysis of Santos et al. (2015) in Polistes paper wasps suggested that the genus originated in a tropical area, and (2) very few temperate solitary wasps overwinter as adults (Evans and West-Eberhard 1970). Furthermore, Piekarski et al. (2018) recently performed a phylogenomic analysis to explore the relationships between vespid species and consequently proposed the hypothesis that eusociality in Polistinae and Vespinae arose abruptly in a shared ancestor of the two subfamilies with morphological PCB (including size differences) as well as physiological PCB. They suggested that the abrupt appearance of eusociality could have arisen from daughter subfertility (West-Eberhard 1975; Craig 1983; first-brood daughters are likely to be malnourished because only the mother nurses the offspring) in colonies composed of a mother and daughters. The hypothesis of abrupt appearance of castes with morphophysiological differences is fascinating and somewhat surprising to researchers of social hymenopterans. One way to better address this hypothesis is to determine caste-fate determination mechanisms using a range of different species of vespid wasps with varying levels of sociality; in particular, this strategy can elucidate how and when PCB occurred and was made stronger during the evolution of eusociality in vespid wasps.

Temperate paper wasps are suitable animals for exploring the origin of workers because the castes, particularly gynes (foundress-queen candidates of the next spring), are distinguishable by physiological traits related to diapause even though they do not have morphological differences. The females that prepare for overwintering, i.e. store lipids, are classified as diapausing females (i.e. gynes) (Eickwort 1969; Strassmann et al. 1984; Toth et al. 2009; Yoshimura and Yamada 2018a), while those that do not are classified as nondiapausing females, including workers, replacement queens, and mid-season foundresses (Strassmann 1981; Page et al. 1989). Moreover, gynes do not develop mature ovaries before overwintering, while nondiapausing females do so before overwintering to perform oviposition under certain conditions (Haggard and Gamboa 1980; Toth et al. 2009), such as when they are dominant or leave their natal nests. First-brood females of some species mate with early males and found new colonies without overwintering (Strassmann 1981; Page et al. 1989; see also Liebert et al. 2004). Some first-brood females of P. jokahamae may become mid-season founders due to nest collapse or replacement queens due to the disappearance of the foundress queens; the former case is rare according to our observation. However, these females are considered to lay only male eggs because very few males are found in the first brood and the females have no chance to mate (see Miyano 1991; Yoshimura et al. 2019); that is, they cannot become true queens and are often classified into the worker caste (O’Donnell 1998). Hence, whether a P. jokahamae individual enters diapause is strongly linked to whether the individual becomes a queen.

Most researchers consider the caste fate of temperate paper wasps to be finally determined by cues received after emergence, even though PCB may occur during the immature stage (Berens et al. 2015; Judd et al. 2015; Judd 2018; Yoshimura and Yamada 2018b). Such PCB is induced by the nutritional level, vibration, and photoperiod (O’Donnell 1998; Hunt 2006; Hunt et al. 2007; Jeanne and Suryanarayanan 2011; Jandt et al. 2017; Yoshimura and Yamada 2018b). The determinants of caste after emergence include the presence of the queen and broods and the photoperiod (Bohm 1972; Solís and Strassmann 1990; Reeve et al. 1998; Tibbetts 2007; Judd 2018; Yoshimura and Yamada 2018b).

Regarding the effects of photoperiod on caste-fate determination, Bohm (1972) revealed the existence of PCB (i.e. preimaginal diapause-related bias) by exposing females of the paper wasp Polistes metricus from different broods to short or long days. The first brood had no PCB, while the other broods exhibited PCB toward gynes. Our previous study (Yoshimura and Yamada 2018b) revealed that the photoperiod during the adult stage is a cue for caste-fate determination in Polistes jokahamae: when females experience long and short days during the adult stage, they exhibit the physiological traits of nondiapausing and diapausing forms, respectively. However, there were no differences in the effect of photoperiod between different-brood females, probably because the colonies in that study were collected before the emergence of the first brood, placed under constant temperatures and a controlled photoperiod, and supplied with enough food. In contrast, the first-brood P. jokahamae adults emerging from nests collected in the field emerge with a bias toward workers: they are small and thin and emerge with lower lipid stores (Yoshimura and Yamada 2018a). Thus, when P. jokahamae females are exposed to similar experimental procedures to those of Bohm (1972), they may exhibit PCB; in particular, the first brood may do so toward workers, that is, individuals with nondiapausing physiological characteristics (Table 1).

Table 1 Summary of predictions and experimental results for preimaginal caste-related biasing (PCB): emerging adults were exposed to long days (LD 16:8) or short days (LD 12:12) in the experiment
Full size table

Here, we aimed to elucidate the existence of PCB (i.e. preimaginal diapause-related bias) through the examination of the effects of photoperiod during the adult stage on caste-fate determination using females from different broods of P. jokahamae. When all adults of a brood exhibited mature eggs and low lipid stores under both short and long days, the brood was assumed to exhibit a strong preimaginal bias toward the nondiapausing form, which usually becomes a worker (Table 1). In contrast, when all adults of a brood refrained from develop ovaries and had large lipid stores under both short and long days, the brood was assumed to present a strong preimaginal bias toward the diapausing form (gynes). When all females exhibited egg maturation and low lipid stores under long days and all females refrained from developing ovaries and had large lipid stores under short days, the brood was assumed to exhibit no PCB or quite weak PCB (Table 1). In addition, we analyzed the effects of the body size of adults on caste-fate determination because smaller females were less likely to prepare for diapause in our previous experiments (Yoshimura and Yamada 2018a, b).

Materials and methods

Study species

Polistes jokahamae is distributed throughout Japan except for in Hokkaido and is particularly common in the southern area of Japan. Foundresses start to solitarily found nests around late April in Mie, Japan, after they have overwintered. Daughters start to emerge from late May to mid-June and usually continue to emerge until mid- to late August (Yoshimura et al. 2019). Males emerge mainly in August. Gynes mate from late October to early November and then enter diapause.

Collection of colonies and categorization of broods

A total of 38 colonies with pupae were collected from May 23 to August 18, 2016, and from May 22 to August 21, 2017, in Tsu, Mie, Japan (Table 2). The collected nests were kept at 25 °C under photoperiod conditions simulating changes in the natural day length, which was defined as 1 h longer than the period from sunrise to sunset. Adults emerging from cocoons that were present when the nests were collected were used for the experiments, providing a total of 191 females for the analysis (Table 2).

Table 2 Sample sizes used for statistical analysis
Full size table

The females in the nests were divided into the first, second, and third broods. The first brood was nursed only by the foundress. The second brood was nursed by the foundress queen and first-brood females and emerged before male emergence. The third brood was also nursed by the foundress and workers but emerged at the same time as or later than the males. The first brood emerged from late May to mid-June, the second brood emerged from mid-June to July, and the third brood emerged mainly in August.

Treatments

Newly emerged females were removed from the natal nests immediately after emergence and placed individually in transparent plastic cups (12 cm diameter × 6 cm deep) for 2 weeks under a 16:8 h LD or a 12:12 h LD at 25 °C, representing day lengths similar to those at the time of the summer solstice (1.5 h longer than the period from sunrise to sunset) and in mid-October (1 h longer than the period from sunrise to sunset), respectively. One third-to-fifth-instar silk moth (Bombyx mori) larva was placed in each rearing cup with water and honey. On emergence days (before food and water were supplied), the head width of the adult wasps was measured using calipers with a precision of 0.01 mm. The water and honey were renewed once per week, and the moth larva was replaced when it was injured.

After being reared for 2 weeks, the wasps were dissected under a microscope to examine whether they exhibited mature eggs. In addition, the lipid stores in the gaster were estimated by measuring the difference in dry mass before and after lipid extraction using diethyl ether as described by Tibbetts et al. (2011). The dry mass was measured using an electronic balance with a precision of 0.0001 g.

Data analysis

First, we compared the body size (head width) of the wasps between the two investigation years and between different broods. The analysis was performed using a mixed GLM (General Linear Model) implemented in NCSS (version 11, NCSS Statistical Software, Kaysville, UT, USA). The colony was incorporated in the models as a random factor. The results showed that adult size differed among different broods and between the two years, indicating multicollinearity among the three factors of brood, body size, and year. The values of the variance inflation factor (VIF), which is often used as a measure of the strength of multicollinearity, for each explanatory value were not high (1.08–1.86). However, if a given factor is related to another factor, the factor may be excluded from the minimally adequate model (Zuur et al. 2010), although it has a significant effect on the response variable. Thus, when performing the below analyses, the brood, body size, and year were incorporated separately in the statistical models to avoid multicollinearity. In addition, we compared the emergence day of the wasps in each brood between the two years to test whether a possible yearly difference in the emergence day caused the above yearly difference in the body size and/or a possible yearly difference in egg maturation and lipid stores.

The effects of the following four potential explanatory variables and the two-way interactions between them on egg maturation (proportion of wasps with mature eggs) were analyzed with logistic regression: (1) day length during the adult stage, (2) adult size (head width), (3) brood, and (4) year. Parameters other than adult size were included in the model as categorical variables. We performed a logistic mixed-model analysis using the “lme4” package in R software (version 3.4.3; R Foundation 2017). A logit link function was applied, and the colony was incorporated as a random factor, including a random slope against head width and a random intercept. The significance of each factor was determined by a likelihood ratio test for the models with and without a focal factor. Starting with the interactions in the full models including the day length, one of the parameters of adult size, brood, or year, and their two-way interactions, we tested the significance of each factor using backward stepwise regression analysis. The focal term was removed from the model when it was not significant. We report the P values for the individual terms: those for nonsignificant terms were obtained when the terms were removed, and those for significant terms were obtained when the terms were removed from the minimally adequate model. When a difference was detected between broods, a sequential Bonferroni multiple comparison test (Holm 1979; Rice 1989) was performed to identify the pairs exhibiting statistically significant differences using the “multcomp” package in R software.

The index of relative lipid stores (IRL), which was defined as (lipid stores)/(head width cubed) (Yoshimura and Yamada 2018a), was used in the analysis to control for body size. The factors influencing the IRL were analyzed with a mixed GLM incorporating the day length and one of the parameters of brood, body size, year, or egg maturation (absence or presence of mature eggs) as fixed factors using the “lme4” package in R software, as the factor of egg maturation was also related to the other three factors, causing multicollinearity (VIF = 1.28). The colony was incorporated as a random factor. A logarithmic transformation was applied to the response variable (IRL) to ensure that the random errors conformed to a normal distribution.

Furthermore, we analyzed the effects of the emergence days of individual wasps on egg maturation and lipid stores separately for each of the three broods. The photoperiod and emergence day were incorporated as fixed factors, and the colony was incorporated as a random factor.

Results

Body size and emergence day

Body size was significantly larger in 2016 than in 2017 (Table 3; F1, 33.6 = 4.6, P = 0.039) and differed significantly between broods (F2, 31.8 = 18.0, P < 0.001), with no significant interaction (F2, 29.4 = 0.3, P = 0.773). The body size of the second and third broods was larger than that of the first brood in 2016 (Table 3), but the difference was not significant in 2017: note that this yearly difference may be due to the low numbers of colonies collected in 2017 (Table 2). The emergence days did not differ between different years (F1, 13.9 = 0.1, P = 0.737 for first brood; F1, 16.9 = 0.6, P = 0.450 for second brood; F1, 1.9 = 8.5, P = 0.109 for third brood).

Table 3 Body size (head width) of emerging females and emergence day in each brood
Full size table

Egg maturation

The year and day length during the adult stage had independent effects on the proportion of females with mature eggs (\({\chi }_{1}^{2}\) = 24.5, P < 0.001 for day length; \({\chi }_{1}^{2}\) = 7.0, P = 0.008 for year): the proportion was higher under long days than under short days and was higher in 2017 than in 2016 (Fig. 1). The interactions of day length with brood and body size were significant (\({\chi }_{2}^{2}\) = 9.1, P = 0.010 for the former interaction;\({\chi }_{1}^{2}\) = 6.2, P = 0.013 for the latter). To explore the mechanisms underlying the interactions, statistical analysis was performed separately for different broods and for short and long days during the adult stage.

Fig. 1
figure 1

Effects of brood and day length during the adult stage on egg maturation (proportion of female adults with mature eggs) in 2016 and 2017. The bar heights represent the proportions of female adults with mature eggs; the bars are classified by broods. The numbers above the bars indicate sample sizes. Second- and third-brood females were more likely to develop ovaries under long days, but first-brood females were not (see Table 4)

Full size image

The analysis of each brood revealed that long days induced higher levels of egg maturation compared with short days in the second and third broods (Table 4; Fig. 1) but not in the first brood. Egg maturation was unrelated to body size in every brood (Table 4), but smaller wasps were more likely to exhibit mature eggs than larger ones when all broods were analyzed together under short days (Table 5; Fig. 2). The brood had a significant effect under both short and long days (Table 5, the third brood was less likely to develop ovaries), although the multiple comparison test did not detect significance between any brood pairs. These analyses suggested the possibility that the body-size difference caused the brood difference under short days. The year had a significant effect on egg maturation in the first brood (Table 4).

Table 4 Statistical results for the effects of the day length, body size, and year on the proportion of adults with mature eggs in each brood
Full size table
Table 5 Statistical results for the effects of the brood and body size on the proportion of adults with mature eggs under short and long days during the adult stage
Full size table
Fig. 2
figure 2

Effects of body size (head width) and day length during the adult stage on egg maturation (proportion of female adults with mature eggs) in 2016 and 2017. The bar heights represent the proportions of female adults with mature eggs; the bars are classified by head width. The numbers above the bars indicate sample sizes. ND, no data. Smaller wasps were more likely than larger wasps to develop ovaries under short days (see Table 5)

Full size image

The emergence day influenced egg maturation among the second-brood females under long days (\({\chi }_{1}^{2}\) = 9.3, P = 0.002, not presented here in tables or figures): females that emerged earlier were more likely to develop ovaries. However, such a difference was not detected in the first or third brood (\({\chi }_{1}^{2}\) = 0.2, P = 0.680 for first brood; \({\chi }_{1}^{2}\) = 2.6, P = 0.109 for third brood).

Lipid stores

The day length during the adult stage affected the IRL without interacting with any other factors (Table 6): the IRL was higher under short days than under long days (Fig. 3), was higher in 2016 than in 2017, and was positively correlated with adult size (Fig. 4). The IRL was significantly lower for the first brood than for the third brood (sequential Bonferroni multiple comparison test, P < 0.05; Fig. 3). Additionally, the IRL did not differ significantly between females with and without mature eggs (Table 6; Fig. 3), which differed from the expectation that it would be lower for females with mature eggs (Yoshimura and Yamada 2018b).

Table 6 Statistical results for the effects of the day length, brood, body size, year, and egg maturation on the IRL [(lipid stores)/(head width cubed)]
Full size table
Fig. 3
figure 3

Effects of the brood and photoperiod during the adult stage on the IRL [(lipid stores)/(head width cubed)] among females without and with mature eggs in 2016 and 2017. Data are presented as the mean and SE. The numbers in the bars indicate sample sizes. ND no data. The IRL was higher under short days than under long days and was lower for the first brood than for the third brood (see Table 6 and the text)

Full size image
Fig. 4
figure 4

Relationship between the IRL [(lipid stores)/(head width cubed)] and body size (head width) under short and long days. IRL values were logarithmically transformed. Lines represent the best-fit linear model estimated for short and long days. The IRL was higher under short days than under long days and was positively correlated with adult size (see Table 6)

Full size image

The emergence day influenced the IRL among second-brood females (\({\chi }_{1}^{2}\) = 5.7, P = 0.017; not presented here in tables or figures) independently of the effect of day length (interaction: \({\chi }_{1}^{2}\) < 0.001, P = 0.987): later-emerging females were more likely to store lipids. Such an influence was not found for the first or third brood (\({\chi }_{1}^{2}\) = 0.2, P = 0.680 for first brood; \({\chi }_{1}^{2}\) = 2.6, P = 0.109 for third brood).

Discussion

Taken together with the results of our previous study (Yoshimura and Yamada 2018a), the present findings suggested that the first brood emerged with a strong physiological bias toward the nondiapausing form, while the second and third broods emerged with no bias or an easily reversible bias regarding diapause at the adult stage (see Table 1). In the field, the first-brood females become workers in the presence of the healthy colony, and many second-brood P. jokahamae adults also become workers (Tsuchida 1991; Yoshimura et al. 2019). Thus, workers on the nest include those with and without a physiological bias toward the nondiapausing form at emergence. Moreover, the present results repudiate the idea that gynes (diapausing females) always emerge after a certain day, which is estimated to be around the first emergence day of males in many species (Table 1; Metcalf 1980; Suzuki 1986; Toth et al. 2009). Reeve et al. (1998) observed that many first-brood females of Polistes fuscatus left their natal nests within a few days of emergence, probably to overwinter and become foundresses the following spring. However, such early dispersal from the natal nest was not observed among first-brood females of P. jokahamae colonies under field and semi-field conditions (Yoshimura et al. 2019). In addition, the present study suggests that first-brood females of P. jokahamae do not overwinter, even when they are separated from the nests.

The first brood emerged with a strong bias toward the nondiapausing form in the present study. However, this phenomenon was not found in laboratory-reared colonies (Yoshimura and Yamada 2018b). Given that the colonies were collected before the emergence of the first female and were reared in temperature- and photoperiod-controlled rooms with sufficient food (honey and moth larvae) in the previous experiments, there were two possible explanations for the difference: (1) the preimaginal diapause-determining factor is gradual changes in day length during the immature stage, and the laboratory-reared first brood did not experience such changes; or (2) the preimaginal diapausing-determining factor is body size or physiological characteristics related to body size. The present analysis did not reveal which explanation is correct because the first-brood females were smaller than the other-brood females, causing a multicollinearity problem. The former explanation was supported by the fact that females that emerged later were more likely to enter diapause in the second brood, while the latter explanation was supported by the first brood being smaller than the other broods in the present experiments but not in the previous experiments (Yoshimura and Yamada unpubl.). Specific experiments will be required to allow a conclusion to be reached: both of the above-mentioned factors may be effective predictive variables. It would be fruitful to determine the response of emerging adults to the photoperiod after placing larvae and/or pupae under treatments, such as different food supplies or day lengths. Hunt and colleagues (Rossi and Hunt 1988; Karsai and Hunt 2002; Hunt and Amdam 2005; Judd et al. 2015) investigated how food availability during the larval stage influenced caste fate. The experimental results were complex. For example, restricted food availability produced females that were more likely to exhibit mature eggs (biased toward the nondiapausing form) but that emerged with larger lipid stores (biased toward the diapausing form) (Judd et al. 2015). These authors did not consider the effects of the photoperiod and reared experimental wasps under constant long days (an LD of 16:8 h). Consideration of the photoperiod might produce clearer results.

The effects of day length on the caste fate of P. jokahamae definitely contrasted with those found in P. metricus (Bohm 1972). In P. metricus, the second and third broods emerged with PCB toward gynes (diapausing form), while the first brood emerged with no bias or easily reversible bias toward workers (nondiapausing form). One possible reason for the difference between P. jokahamae and P. metricus is that the first brood emerges later in P. metricus than in P. jokahamae (Bohm 1972; the first brood of P. metricus emerged in mid- and late June). The pattern of the appearance of PCB found in different broods of P. metricus appears to be adaptive for paper wasps with shorter periods of colony activity. The second and third broods are thought to always enter diapause in the field due to strong PCB. The first brood may also enter diapause when it emerges later than usual due to cooler temperatures. It should be noted that P. metricus is widely distributed from Florida to New York State in the US (Buck et al. 2008). The populations in southern areas may exhibit similar patterns of PCB appearance to that in P. jokahamae. Interestingly, no workers are present in Polistes biglumis nests found in alpine areas, while workers are present in warmer mountain areas (Fucini et al. 2009, 2014). This situation may represent an extreme case of adaptation to a shorter activity period.

The second brood of P. metricus develops ovaries when JH is applied (Bohm 1972), suggesting that egg maturation is regulated by JH; such a system is widespread in insects, including other paper wasps, such as Ropalidia marginata and Polistes dominulus (Robinson and Vargo 1997; Jandt and Toth 2015; Kapheim 2017). Tibbetts and Izzo (2009) and Tibbetts et al. (2011) reported that the relationship between egg maturation and the level of JH changes depending on the social status and body size of the focal individuals in P. dominulus. This suggests the possibility that the relationship also changes depending on day length during the adult stage. However, our team (Yoshimura, Yamada, and Sasaki unpubl.) suggested another possibility. We have revealed that the brain levels of two biogenic amines, tyramine and dopamine, were positively correlated with egg maturation in P. jokahamae and that the brain level of tyramine responded to photoperiod during the adult stage, but that of dopamine did not. This suggests that tyramine regulates egg maturation in response to photoperiod. However, the photoperiod may also influence egg maturation through the JH signaling pathway (see Miki et al. 2020) in P. jokahamae. This is partially supported by the fact that JH regulates the level of dopamine in Polistes chinensis antennalis (Tsuchida et al. 2020). Unfortunately, the level of JH was not determined in the above study. Whether JH is involved in caste-related physiological responses to photoperiod in paper wasps requires further study.

Piekarski et al. (2018) hypothesized that the shared ancestor of Polistinae and Vespinae had workers with PCB toward workers at emergence, but the present study suggests the possibility that the ancestor had workers who emerged without PCB as well. The caste fate of the females emerging without PCB would be determined mainly by colony-related factors during the adult stage. Since the ancestor is considered to have been distributed in a tropical area (Santos et al. 2015), environmental factors, such as photoperiod and temperature, do not appear to be important. Social situations (e.g. subordinate females are less likely to get food) are as likely to change behavioral and physiological caste-related characteristics as body size. This is because the two kinds of changing process appeared to occur without genetic changes and to be closely related to the reproduction-related responses conserved in most insects, including solitary wasps: females are more likely to develop ovaries and produce more eggs when they are larger and greater amounts of nutrients are available (Cowan 1981; Honěk 1993; Tibbetts et al. 2013; Kapheim 2017).

We observed that many of the wasps categorized as females without mature eggs exhibited nearly mature oocytes under long days. The IRL of all broods, including the first brood, was higher under short days than under long days. In addition, the IRL was not higher among females without mature eggs than among females with mature eggs. Our previous study (Yoshimura and Yamada 2018a) revealed that the IRL was approximately 0.14 mg/mm3 among second- and third-brood females just after emergence. This value was similar to or higher than the IRL among the females reared for 2 weeks in the present study. This finding suggests that most females, including diapausing females, show no increase in the IRL for 2 weeks after emergence, although some females use some of their lipid stores for egg maturation and suppress the increase in the average IRL. Yoshimura (unpubl.) discovered that females collected in the field in early winter exhibited an IRL of approximately 0.28 mg/mm3. Our preliminary experiment (Yoshimura and Yamada unpubl.) in which females emerging from nests collected in the field were reared for 4 weeks according to the same procedures used in the present study showed that many second- and third-brood females without mature eggs exhibited an IRL of approximately 0.28 mg/mm3 under short days and that up to 80% of the first-brood females exhibited mature eggs under long days. These findings suggest that most of the females that will finally enter diapause show no substantial increase in lipid stores for 2 weeks after emergence. The females appeared to decide to enter diapause after spending a few weeks receiving cues related to caste fate (i.e. diapause) determination, including day length and colony conditions.

The proportion of female adults with mature eggs in the first brood was greater in 2017 than in 2016. Body size had no effect on the proportion in the first brood. The emergence days did not differ in different years with no effect on the proportion. Thus, the yearly difference in the proportion was considered to be caused by other factors. One possible factor was the yearly difference in the quality and quantity of food that the individuals received at different stages of the larval period. This may have caused the difference in the amount of JH released and the physiological differences among the emerging females between the two years (Kapheim 2017). However, it should be noted that the difference between years might decrease or even disappear if the emerging females were reared for a longer period. This is because our preliminary experiment, which was mentioned in the previous paragraph, discovered that when first-brood females were reared for 4 weeks under long days, the proportion of females with mature eggs reached approximately 80% in each of the two years studied.

The present study strongly suggests that the rearing of different-brood adults under different photoperiods is an effective way to reveal the strength of PCB. The same experimental procedures are expected to be performed in many other paper wasp species. Then, the following further questions should be raised to explain the evolutionary process of eusociality: how photoperiod interacts with other PCB-generating factors, such as colony-related factors, and how photoperiod induces physiological changes related to PCB. We have already started experiments to solve these questions and to determine the effects of photoperiod on caste-fate determination using other paper wasp species. We firmly believe that many novel findings are waiting to be disclosed.