Effects of Self-Superparasitism and Temperature on Biological Traits of Two Neotropical Trichogramma (Hymenoptera: Trichogrammatidae) Species

Abstract

It is common for a female trichogrammatid to lay more than one egg per host, a phenomenon known as self-superparasitism, which exposes her offspring to intraspecific, intrinsic competition (IIC) with its own siblings. Information about how often self-superparasitism occurs and how IIC interacts with abiotic factors is rare, especially regarding the Neotropical Trichogramma species. Here we determined the frequency of self-superparasitism in Trichogramma atopovirilia Oatman & Platner ( Ta ) and T. pretiosum Riley ( Tp ), and the effects of IIC and temperature on the sex ratio, egg-to-adulthood period, and survivorship of both species’ offspring. Individual females were offered eggs of Spodoptera frugiperda (J.E. Smith) for 30 min. A group of parasitized hosts was then dissected for determination of the self-superparasitism frequency, while another group was incubated at 15, 18, 21, 24, 27, 30, and 33°C. High rates of self-superparasitism were found in both Ta (0.55 ± 0.07) and Tp (0.62 ± 0.06). IIC interacted with temperature decreasing Ta and Tp ’s survivorship, lengthening the egg-to-adulthood period in Tp and shortening it in Ta , and balancing Ta ’s sex ratio. Based on survivorship rate, Ta and Tp could not be differentiated if their immatures develop in absence of IIC. However, in its presence, Tp was 3 × more likely to survive than Ta at 33°C, while at 15°C Ta survived 2× better than Tp . These results show that self-superparasitism can be very common in both Ta and Tp , and that its effects on the species’ biological traits and competitiveness strongly depend on the IIC–temperature interaction.

Resumo (Portuguese)

É comum que fêmeas de tricogramatídeos coloquem dois ou mais ovos por hospedeiro, fenômeno conhecido como auto-superparasitismo, o qual gera competição intrínseca intraespecífica (CII) entre irmãos. Informações sobre frequência de auto-superparasitismo e como CII interage com fatores abióticos são raras, sobretudo em relação às espécies neotropicais de Trichogramma . Neste estudo, determinaram-se a frequência de auto-superparasitismo em Trichogramma atopovirilia Oatman & Platner ( Ta ) e T. pretiosum Riley ( Tp ); assim como os efeitos das CII e temperatura na razão sexual , período ovo-adulto e sobrevivência da prole. Ovos de Spodoptera frugiperda (J.E. Smith) foram oferecidos a fêmeas individualizadas por 30 min. Parte dos hospedeiros submetidos ao parasitismo foi dissecada para determinação do auto-superparasitismo, enquanto outra parte foi incubada a 15, 18, 21, 24, 27, 30 e 33°C. Altas taxas de auto-superparasitismo foram encontradas em ambas, Ta (0.55 ± 0.07) e Tp (0.62 ± 0.06). CII interagiu com temperatura diminuindo a sobrevivência em Ta e Tp ; alongando o período ovo-adulto em Tp ou encurtando-o em Ta ; e balanceando a razão sexual em Ta . Baseando-se na sobrevivência, Ta e Tp não se diferenciaram quando seus imaturos se desenvolveram na ausência de CII. Contudo, em sua presença, a probabilidade de Tp sobreviver a 33°C foi 3× superior à de Ta ; enquanto a 15°C Ta apresentou 2× mais chances de sobreviver do que Tp . Estes resultados mostram que auto-superparasitismo pode ser muito comum em Ta e Tp , e que os seus efeitos sobre as características biológicas e competitividade destas espécies dependem fortemente da interação CII-temperatura.

fall armyworm, intraspecific competition, intrinsic competition, Trichogramma atopovirilia, Trichogramma pretiosum

Superparasitism, which occurs when a female parasitoid lays a new clutch of eggs in or on an already parasitized host, is a common phenomenon in insect parasitoids ( Van Alphen and Visser 1990 , Montoya et al. 2012 ), especially in the genus Trichogramma (Hymenoptera: Trichogrammatidae) ( Narayanan and Chacko 1957 , Van Dijken and Waage 1987 , Moreira et al. 2009 , Bueno et al. 2010 , Shoeb and El-Heneidy 2010 ).

When superparasitism occurs, the new clutch of eggs may be laid either by the same female that performed the first oviposition or by a conspecific female, events respectively known as “self-” and “conspecific” superparasitism ( Van Dijken and Waage 1987 ). In either case, the immatures will engage in intraspecific, intrinsic competition (IIC), sharing the resources of a single host with no chance of escaping and looking for another host (reviewed by Harvey et al. 2013 ).

IIC has been shown to affect a wide range of physiological, morphological, and behavioral characteristics of parasitoids, including size, fecundity, ability to mate, sex ratio, egg-to-adulthood period, locomotion, and survivorship ( Narayanan and Chacko 1957 , Suzuki et al. 1984 , Wajnberg et al. 1989 , Bai and Mackauer 1992 , Harvey et al. 1993 , Moreira et al. 2009 , Montoya et al. 2012 ). Given that most of these effects negatively interfere with the parasitoid colonization process, IIC can play a critical role on the success of biological control programs ( Narayanan and Chacko 1957 , Wajnberg et al. 1989 ). However, studies of the frequency at which it occurs, as well as how it interacts with abiotic factors are still scarce, especially in regards to the Neotropical Trichogramma species.

Temperature is often the most detrimental abiotic factor influencing insect populations and distribution ( Damos and Savopoulou-Soultani 2012 ). From a macro perspective, host eggs suffer seasonal and regional temperature variation, and might be impacted by climate change ( Easterling et al. 2000 , Bale et al. 2002 , Hance et al. 2007 , Robinet and Roques 2010 ). From a micro perspective, given that most host eggs are laid on leaves and that temperature of different leaves from a single plant can be substantially distinct (±3°C) even when they are at similar height, orientation, shading, and color ( Potter et al. 2009 ), host eggs might be subjected to a range of temperatures even at plant level and throughout the day. Temperature has been proven to affect several life history traits of egg parasitoids, including of Trichogramma ( Melo et al. 2007 ; Bueno et al. 2008 , 2009 , 2010 ). Trichogramma atopovirilia ( Ta ) and T. pretiosum ( Tp ) are two Neotropical trichogrammatid species with potential for controlling the fall armyworm, Spodoptera frugiperda (Lepidoptera: Noctuidae), in augmentative biological control programs, in Brazil ( Parra and Zucchi 2004 , Beserra et al. 2005 ). Eggs of the fall armyworm are laid on both faces of maize ( Zea mays L.) leaves, from the bottom to the top of the plants ( Beserra et al. 2002 ), thus being exposed to temperature variation. Ta and Tp practice self-superparasitism ( Beserra and Parra 2004 , Bueno et al. 2010 ), but its frequency and effects on the biology of each species are still unknown. Here we studied 1) the frequency of self-superparasitism of Ta and Tp on eggs of the fall armyworm; and 2) the effects of IIC and temperature on three biological traits of both species.

Materials and Methods

Insects

Colonies of the fall armyworm, Ta and Tp were kept in laboratory (24 ± 1°C, 70% RH, and a photoperiod of 14:10 [L:D] h) from insects collected in a maize field, as described by Da Silva and Parra (2013) and DaSilva et al. (2015) .

Parasitism Procedures

Fall armyworm eggs were carefully detached from fresh (<12 h) egg masses with a fine camel-hair brush and placed in a Petri dish (9 cm Ø). Then, seven eggs were randomly chosen and grouped in the center of a piece (1 by 1 cm) of paper (obtained from a cage where adults of S. frugiperda were reared, thus carrying host’s chemical cues). Distilled water was used to attach the eggs to each other and to the paper. The seven eggs were set with their micropila facing upwards, as S. frugiperda naturally position them. The egg-carrying paper was placed into a test tube (1.2 by 7 cm) closed with a PVC film containing a single female parasitoid (1–3 d old, mated, honey-fed, no oviposition experience), and as soon as the female encountered the grouped eggs, we started a 30-min countdown (long enough for parasitism of seven eggs in preliminary tests) at room conditions (24 ± 2°C, 70 ± 10% RH, and a photoperiod of 14:10 [L:D] h). Seventy different individual females of each species performed the parasitism, totaling 490 host eggs (70 groups of 7 host eggs) submitted to parasitism.

Frequency of Self-Superparasitism and Incubation Procedures

Self-Superparasitism

After 30 min the female parasitoids were eliminated and 51 of the host eggs submitted to parasitism by each species were randomly chosen and dissected under a microscope (40× magnification) in order to estimate the frequencies of unparasitized, singly parasitized, and superparasitized hosts (i.e., host eggs containing 0, 1, or 2+ parasitoid eggs), as well as the mean number of parasitoid eggs laid per host egg.

Incubation

After 30 min the female parasitoids were eliminated and 350 of the host eggs submitted to parasitism by each species were randomly chosen and individualized with a moistened brush on a new piece of paper square. Individual host eggs were then inserted in test tubes (as previously described) and distributed in incubators at seven different temperatures (15, 18, 21, 24, 27, 30, and 33 ± 1°C, 70 ± 10% RH, and a photoperiod of 14:10 [L:D] h), totaling 50 host eggs/temperature/species. The two egg parasitoid species and the seven temperatures were tested simultaneously. Parasitoid emergence was checked twice a day (hours 2 and 10 of the photophase) for 50 d under a stereoscopic microscope, and the number/sex of emerged adults as well as the collapsed hosts were recorded.

Impact of IIC and Temperature on Survivorship Rate

Following parasitoid emergence, we calculated the observed rates of host eggs that generated no (“0”), singly (“1”) or multiple (“2+”) adults, which were contrasted to their respective expected rates of emergence (= rates of hosts containing 0, 1 or 2+ parasitoid eggs, obtained according to the description in the previous topic), at each temperature. When the observed and expected rates matched ( P  > 0.05) at a given temperature, we inferred that IIC had no impact on the parasitoid survivorship rate. On the other hand, when both rates did not match ( P  ≤ 0.05), we concluded that the parasitoid mortality was significant and hence IIC impacted the survivorship at that specific temperature.

Impact of Species and Temperature on Survivorship Rate

The first step of this experiment was to identify, amongst all the emerged adults, which ones developed in presence or absence of IIC. For such, we followed the logical steps described in the Fig. 1 . Initially, we spotted the host eggs that gave rise to 2 or more (“2+”) adults, which undoubtedly indicated that such adults developed in presence of IIC. Unfortunately, we could not conclude that every adult that emerged alone developed in absence of IIC because such adult may had developed with a competitor sibling that failed to emerge (i.e., the sibling died in the embryonic, larval, or pupal stage); hence, such adult would have developed in presence of IIC even tough it emerged alone. We then found that the “observed – expected” contrasts performed in the previous topic could be useful for identifying the singly developed adults. Firstly, if the observed and expected matched ( P  > 0.05) simultaneously in the “1” and “2+” adults, we concluded that the singly emerged adults developed in absence of IIC ( Fig. 1 ). Alternatively, if the observed and expected rates matched in “1” but did not match in “2+” while the observed rate of no emergence (“0”) was higher ( P  ≤ 0.05) than its expected rate, we also concluded that the singly emerged adults developed in absence of IIC ( Fig. 1 ). This IIC identification system worked because the rates of “0”, “1,” and “2+” adults at each temperature-species combination were interdependent. For example, a decrease in “2+” had to cause an increase in “0” and/or “1,” and an increase in “0” had to be caused by a decrease in “1” and/or “2+.” The occurrence of mistakes in this identification system could be easily spotted as shown in the Fig. 1 .

Fig. 1.

Depiction of the system employed to identify which adults of Trichogramma atopovirilia and T. pretiosum developed in absence or presence of IIC following parasitism of eggs of S. frugiperda (15–33 ± 1°C, 70 ± 10% RH, and a photoperiod of 14:10 [L:D] h).

Following the IIC identification, we calculated the survivorship rates of the offspring that developed in the presence and absence of IIC. Generalized linear models (GLM) of logistic regression finally measured the simultaneous impact of parasitoid species and temperature on the survivorship rate for each IIC status (Absence of IIC: N  = 17–40; Presence of IIC: N  = 30–44, depending on the temperature–species combination), as described in the section “Statistical analysis.”

Impact of IIC and Temperature on Egg-to-Adulthood Period and Sex Ratio

Following the IIC identification, it was calculated the egg-to-adulthood period (d) and the sex ratio (number of emerged females/total number of emerged adults) of the adults that developed in the presence or absence of IIC. The impact of IIC and temperature on such life history traits was finally measured for each parasitoid species by a two-way ANOVA ( Ta : N  = 10–32; Tp : N  = 6–33, depending on the temperature–IIC status combination), as described below.

Statistical Analysis

Contrasts between Ta and Tp (number of parasitoid eggs per host egg), and between expected and observed rates of emergencies were performed by the Mann–Whitney two-tailed U test (α = 0.05). Generalized linear models (GLM) of logistic regression were performed in order to assess the simultaneous effects of parasitoid species and temperature on the survivorship rate of each IIC status (presence and absence) (α = 0.05). In case of significance, multiple comparison tests (Tukey’s, α = 0.05) were applied in the same logistic regression model in order to identify the significant differences (α = 0.05). Two-Way ANOVA (α = 0.05) was employed to assess the simultaneous effects of IIC status and temperature on the egg-to-adulthood period and sex ratio after confirmation of normality and homoscedasticity of variances. In case of significance, the Tukey’s test was applied in order to identify the significant differences (α = 0.05). GLM procedures and following multiple comparisons were performed in R (R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/ ). Mann–Whitney, two-way ANOVA and following Tukey’s tests were ran on GraphPad Prism version 6.0f for Mac OS X (GraphPad Software, La Jolla, CA, www.graphpad.com ).

Results

Frequency of Self-Superparasitism

In Ta we found 7 (14%), 16 (31%), and 28 (55%) host eggs containing, respectively, 0, 1 and 2 parasitoid eggs. In Tp , we found 8 (15%), 12 (23%), 24 (47%), 6 (13%), and 1 (2%) host eggs containing, respectively, 0, 1, 2, 3, and 4 parasitoid eggs per host egg ( Fig. 2 a). Such values produced an average of 1.41 ± 0.10 and 1.61 ± 0.13 parasitoid eggs per fall armyworm egg, respectively, for Ta and Tp ( Fig. 2 b), with no significant difference between both species ( N  = 51, U  = 1148, P  = 0.27).

Fig. 2.

( a ) Frequency (mean ± SEM) of host eggs ( S. frugiperda ) parasitized with 0, 1, 2, 3, or 4 eggs of Trichogramma atopovirilia ( Ta ) or T. pretiosum ( Tp ). ( b ) Number of parasitoid eggs per host egg (mean ± SEM) for Ta and Tp (24 ± 2°C, 70 ± 10% RH, and a photoperiod of 14:10 [L:D] h) . Ns = no significant differences (unpaired, two-tailed Mann–Whitney U test, N  = 51, U  = 1148, P  = 0.27).

Impact of IIC and Temperature on Survivorship Rate

No impact of IIC on the survivorship of Ta was found from 15 to 30°C ( Fig. 3 a–c). On the other hand, IIC negatively affected the survivorship of Ta at 33°C, when the observed rate was less than half than the expected rate of emergence in multiple emerged adults (“2+”; U  = 793, P  = 0.0002; Fig. 3 c), while the observed rate more than doubled that of the expected rate in no emergence (“0”; U  = 885.5, P  = 0.0016; Fig. 3 a) as a result of the increased mortality in self-superparasitized hosts. IIC negatively impacted the survivorship also of Tp at 15 (“2+”: U  = 831.5, P  = 0.0006; “0”: U  = 863, P  = 0.0006) and 18°C (“2+”: U  = 949, P  = 0.02; “0”: U  = 1038, P  = 0.04; Fig. 3 d and f). No effect of IIC on Tp ’s survivorship was observed from 21 to 33°C ( Fig. 3 d–f).

Fig. 3.

Comparisons between expected and observed emergence rates (mean ± SEM) for 0, 1, and 2+ adult(s) emerged per host egg ( S. frugiperda ). ( a–c ) Trichogramma atopovirilia ; ( d–f ) T. pretiosum (15–33 ± 1°C, 70 ± 10% RH, and a photoperiod of 14:10 [L:D] h). Each observed rate ( N  = 50) was contrasted to its respective expected rate ( N  = 51). Contrasts were performed by the unpaired, two-tailed Mann–Whitney U test (α = 0.05). * P  < 0.05, ** P  < 0.01, *** P  < 0.001.

Impact of Species and Temperature on Survivorship Rate

No effect of parasitoid species, temperature, or their interaction was found for parasitoids that developed singly ( Table 1 ; Fig. 4 a). On the other hand, very strong impacts of both temperature and species–temperature interaction were found for the individuals that developed in presence of IIC ( Table 1 ). In this case, both Ta and Tp survived best at 27°C, although no significant differences were found among 15 and 30°C for Ta , and among 18 and 33°C for Tp ( Fig. 4 b), indicating that the two species differ in the temperature range for best survivorship. Clear-cut differences between both trichogrammatid species were found at 15°C, where Ta survived ca . twice better than Tp ( P  = 0.0069), and also at 33°C, where Tp survived ca . 3× better than Ta ( P  = 0.00001; Fig. 4 b).

Fig. 4.

Impact of species ( Trichogramma atopovirilia and T. pretiosum ) and temperature (15, 18, 21, 24, 27, 30, and 33 ± 1°C) on the survivorship rate (mean ± SEM) of the parasitoid offspring that developed in eggs of S. frugiperda in ( a ) absence or ( b ) presence of IIC (70 ± 10% RH and a photoperiod of 14:10 [L:D] h). Distinct lowercase letters indicate significant differences among same-color columns, while capital letters indicate significant differences between distinct-color columns (generalized linear model of logistic regression followed by multiple comparison tests, α = 0.05). The values displayed at the base of each column indicate the number of replicates.

Table 1.

Generalized linear model (GLM) of logistic regression to evaluate the impact of parasitoid species ( Trichogramma atopovirilia and T. pretiosum ) and temperature (15, 18, 21, 24, 27, 30, and 33 ± 1°C) on the survivorship rate of egg parasitoids reared on S. frugiperda , in absence or presence of IIC (70 ± 10% RH and a photoperiod of 14:10 [L:D] h)

Source of variation	Absence of IIC			Presence of IIC
	df	χ 2	P	df	χ 2	P
Species	1	0.3499	0.5542	1	0.9431	0.3314
Temperature	6	6.3471	0.3854	6	18.7825	0.0045**
Interaction	6	7.2322	0.2999	6	29.5511	<0.0001***

Impact of IIC and Temperature on Egg-to-Adulthood Period and Sex Ratio

IIC status and temperature, as well as their interaction, had very strong impacts on the egg-to-adulthood period of Ta and Tp ( Table 2 ). In both species, the presence of IIC affected ( P  < 0.0001) the length of egg-to-adulthood period specifically at 15°C. However, while in Ta such period was shortened by about 2 d, in Tp it was lengthened in 2 d ( Fig. 5 a, b). A secondary effect of IIC was found in Ta when the temperature range that promoted the fastest development for such species was longer (27–33°C) in the presence of IIC than in its absence (30–33°C; Fig. 5 a). The presence of IIC caused no effect on the egg-to-adulthood period of Ta or Tp from 18 to 33°C.

Fig. 5.

Impact of intraspecific, intrinsic competition status (absence and presence) and temperature (15, 18, 21, 24, 27, 30, and 33 ± 1°C) on the egg-to-adulthood period and sex ratio (mean ± SEM) of Trichogramma atopovirilia ( a and c ) and T. pretiosum ( b and d ), reared on eggs of S. frugiperda (70 ± 10% RH and a photoperiod of 14:10 [L:D] h). Distinct lowercase letters indicate significant differences among same-pattern columns, while capital letters indicate significant differences between distinct-pattern columns (two-way ANOVA followed by the Tukey’s test, α = 0.05). The values displayed at the base of each column indicate the number of replicates.

Table 2.

Two-way ANOVA to evaluate the impact of IIC status (presence and absence) and temperature (15, 18, 21, 24, 27, 30, and 33 ± 1°C) on the egg-to-adulthood period and sex ratio of Trichogramma atopovirilia and T. pretiosum reared on eggs of S. frugiperda (70 ± 10% RH and a photoperiod of 14:10 [L:D] h)

Parameter	Source of variation	T. atopovirilia			T. pretiosum
		df	F	P	df	F	P
Egg-to-adulthood period	IIC status	1	8.89	0.0032**	1	10.48	<0.0001***
	Temperature	6	5069	<0.0001***	6	4550	<0.0001***
	Interaction	6	4.79	0.0001***	6	4.99	<0.0001***
Sex ratio	IIC status	1	18.23	<0.0001***	1	0.02	0.88
	Temperature	6	2.26	0.038*	6	1.29	0.26
	Interaction	6	2.71	0.014*	6	1.51	0.17

Very significant effects of IIC status, temperature, and also of their interaction were also found for sex ratio in Ta ( Table 2 ). At both 15 and 18°C, insects that developed in the presence of IIC exhibited lower ( P  ≤ 0.05) sex ratio than the parasitoids that developed singly ( Fig. 5 c). On the other hand, we found no effect of IIC status, temperature, nor of their interaction on Tp ’s sex ratio ( Table 1 ; Fig. 5 d).

Discussion

After dissecting the host eggs offered to individual egg parasitoid females, we found that self-superparasitism is very common in both Ta (55%) and Tp (62%). Its occurrence in these species is not a novelty, given that other authors have reported the emergence of 1.4 and 1.28 adults of Ta and Tp , respectively, per each S. frugiperda egg ( Beserra and Parra 2004 , Bueno et al. 2010 ), indicating that a certain proportion of host eggs gave rise to two or more adult parasitoids. However, our results are the first to report the frequency at which self-superparasitism occurs in these species. Curiously, both trichogrammatids left unparasitized about 15% of the available host eggs (about 1 out of each 7 hosts), even when females were given enough time to parasitize every host egg. The lack of suitable (unparasitized) hosts has been considered one of the main reasons why female parasitoids practice self-superparasitism ( Van Dijken and Waage 1987 , Darrouzet et al. 2003 ). Hence, not fully utilizing such important resources when given ample opportunity (and carrying a full load of eggs) is an interesting question that still needs to be explored, and whose answer could give insights to the biological control of pests (e.g., how to minimize the rate of unparasitized hosts). On the other hand, the fact that individual females of Ta and Tp frequently performed self-superparasitism even when unparasitized hosts were available suggests that such behavior is an intrinsic characteristic of these species, not necessarily depending on external factors such as the lack of unparasitized hosts, or the risk that later a host will be attacked by a conspecific female ( Van Alphen and Visser 1990 ).

Interestingly, the number of eggs laid per host was similar between Ta (1.41) and Tp (1.61), which along with the fact that the two species showed approximate likelihood to initiate self-superparasitism demonstrates that Ta and Tp suffered similar levels of IIC. However, at 15°C, while IIC shortened the development of Ta by about 2 d, in Tp it was lengthened by 2 d ( Fig. 5 a, b), decreasing the emergence gap from about 3 d to 1 d between the two parasitoid species, and also demonstrating that similar levels of self-superparasitism can cause opposite effects in two closely related egg parasitoid species. Considering that immature parasitoids are often very vulnerable to natural enemies ( Poelman et al. 2012 , Vanaclocha et al. 2013 )—especially egg parasitoids because they are sessile—an extended development period would have a negative impact on a species' competitiveness since it would increase the parasitoid’s exposure to predators and delay its access to unparasitized host eggs. This fact is even more critical for the egg parasitoids of S. frugiperda in Brazil, whose eggs have a wide range of predator species ( Menezes-Netto et al. 2012 ), thus putting the immatures in a super-vulnerable position.

Although IIC caused a decrease on the survivorship of both Ta and Tp , in the latter such effect was observed exclusively at low temperatures (15 and 18°C), while in the former it appeared at the highest temperature tested (33°C) only, thus indicating that the stress caused by temperature or IIC alone is not enough to kill immatures of Ta or Tp , but a combination of IIC and extreme temperatures can be fatal for both species. The contrasts between observed and expected emergencies revealed that every decrease in the rate of multiple emerged adults was followed by a proportional increase in the rate of no emergence. That is, IIC impacted the survivorship of Ta and Tp by the death of all sibling competitors in each host egg, instead of partial death of siblings (i.e., death of all siblings except one). If the latter case was true, a decrease in the observed rate of multiple emerged adults would be obligatorily followed by an increase in the observed rate of singly emerged adults ( Fig. 1 ), which did not happen ( Fig. 3 b, e). This finding shows that the stress caused by the temperature–IIC combination affected all siblings in each host equally, independently of sex and oviposition order. As the observed–expected contrasts suggested ( Fig. 3 b, e), the GLM analysis confirmed that clear differences between Ta and Tp where only visible when IIC and specific temperature conditions (15 or 33°C) were present ( Fig. 4 ). Hence, based on survivorship rate, Ta and Tp are more similar than different egg parasitoid species, although the few discrepancies between them may constitute the borderline that separates a species’ success from failure in extreme temperatures.

Regarding sex ratio, specifically at 15 and 18°C, the emergence of Ta males was very low to zero in singly developed individuals, but IIC increased it to a level verging the balance between sexes ( Fig. 5 c). The high mortality of males in singly parasitized hosts demonstrates that the absence of IIC is detrimental for males of Ta at low temperatures. Literature has reported long ago that differential mortality of sexes in parasitoids can be caused by either temperature ( King 1987 ) or IIC ( Suzuki et al. 1984 ). Nonetheless, the novelty of our study is that we focused on both variables at the same time, which allowed us to go a little further and find that IIC interacts with temperature on the control of differential mortality of sexes such that depending on temperature, the presence of a sibling can dramatically increase the likelihood of a male Ta to survive in comparison to males that developed alone, a mechanism of sex ratio control never cited before. The general concept states that self-superparasitism is advantageous if it elevates the likelihood of an offspring to emerge in case of risk of 1) superparasitism by a conspecific female; 2) fatal attack by the host’s immune system; or 3) if the chances of an offspring to emerge from a host is higher when a female lays two or more eggs in that host ( Viser 1993 , Rosenheim and Hongkham 1996 , Darrouzet et al. 2003 ); all of which are biotic causes. Our findings show, however, that self-superparasitism can also be advantageous when it increases the likelihood of an offspring to emerge at specific abiotic conditions, such as low temperatures (15 and 18°C). It has been demonstrated that for a larva of an egg parasitoid to develop normally it has to consume the whole content of the host egg, because the excess food can increase development time, inhibit the normal pupation and reduce emergence ( Hoffman et al. 1975 , Grenier 1997 ). Moreover, it is known that male trichogrammatids are generally smaller ( Hurlbutt 1987 ) and require less resources to achieve their ideal development than females ( Klomp and Teerink 1966 ). These characteristics could explain why singly developed males of Ta mostly failed in emerging at low temperatures: Their lower metabolism at such abiotic conditions compromised their ability to fully consume the host’s content, thus causing the developing complications recently cited. Further studies should test this hypothesis and also on if the adult female explores such mechanism as a complimentary (secondary) form of sex allocation. In this case, behavioral changes such as avoidance of laying unfertilized (male) eggs at low temperatures—unless the host egg had been already parasitized—would be expected in the parental female in order to avoid male mortality. Ultimately, from a practical perspective, the information on differential mortality of Ta that we bring here may give insights on the development of techniques to avoid the emergence of males for specific experiments (e.g., when a large amount of virgin females is needed) and even for field releases.

The fact that IIC affects the development period, survivorship, and sex ratio of Ta and Tp exclusively at certain temperatures, which were not necessarily the same for both species, clearly shows that each species deals differently with temperature when developing in superparasitized hosts. This was easily observed not only because the temperatures at which IIC had an impact on survivorship were distinct for the two egg parasitoid species ( Fig. 3 c, f), but mainly because 1) such impact of IIC was far stronger in Ta than in Tp ( Fig. 4 b); and 2) Ta ’s sex ratio was also affected by IIC while Tp ’s was not. We found no previous reports of the interaction between IIC and temperature, thus why/how IIC affects Ta and Tp only at certain temperatures still need to be investigated. However, Tp is a far more generalist species than Ta regarding hosts and geographical distribution in Brazil and South America ( Zucchi and Monteiro 1997 , Moreira et al. 2009 , Magalhães et al. 2012 ). Considering host suitability (nutritional value and immune response) as a stressing factor for immature Trichogramma ( Roriz et al. 2006 , Abdel-latief and Hilker 2008 ), one would expect that more generalist species (e.g., Tp )—for being able to interact with a wider range of hosts—evolved to suffer less from host-linked stress than less generalist species (e.g., Ta ). Also, being a geographically widespread species (e.g., Tp ) requires a higher tolerance to temperature variation in comparison to more geographically restricted species (e.g., Ta ). These characteristics help to explain why multiple-developed Tp were generally less affected by IIC-temperature than Ta in our study: Competing siblings had less stressing factors to deal with in Tp than in Ta . From an ecological point of view, when at 33°C IIC increases the mortality of Ta —which self-superparasitized 55% of its hosts—and shortens the life cycle/balances the sex ratio of the species at 15 and 18°C, it is demonstrated that Ta ’s fitness is maximized when IIC occurs at low temperatures and minimized at high temperatures. As opposed to Ta , the fitness of Tp —which self-superparasitized 62% of its hosts—was decreased when IIC occurred at low temperatures (as the mortality and development period were increased at 15 and 18°C). Considering such high potential for self-superparasitism and sex allocation ( Waage and Ming 1984 , Luck et al. 2000 ) there is a chance that natural selection has produced adult Ta and Tp that are able to explore such tools according to temperature in order to maximize their fitness, as previously discussed for sex ratio. Hence, further studies should test if adults of Ta increase their rate of self-superparasitism at low temperatures and avoid it at high temperatures, and if Tp avoids it at low temperatures.

Here it is shown that self-superparasitism can be very common in Ta and Tp —two Neotropical trichogrammatid species. Its primary effect—IIC—interacts with temperature on the control of the offspring’s life-history traits, decreasing Ta and Tp ’s survivorship; lengthening the preimaginal development in Tp and shortening it in Ta ; and balancing Ta ’s sex ratio; in a temperature-dependent manner. Additionally, based on survivorship rate, Ta and Tp cannot be differentiated if their immatures develop in absence of IIC. On the other hand, in presence of IIC, while both trichogrammatids are mostly similar, their few discrepancies have potential to define which species will survive at extreme temperatures. Hence, the fitness gain from self-superparasitism in Ta and Tp is dynamic rather than constant along a temperature range. Since the incubating temperature often varies in both artificial and natural conditions, our findings may help to improve the quality of mass-produced wasps, as well as to better understand the factors that affect the population dynamics of egg parasitoids in field. Further studies should focus mainly on testing 1) if the self-superparasitism rates and sex allocation of Ta and Tp vary with temperature in order to maximize fitness; 2) if host suitability explain why/how IIC affects Ta and Tp only at certain temperatures; 3) how males of Ta benefit from IIC at low (15 and 18°C) temperatures; and 4) if the effects of IIC here observed on siblings are also valid for nonsiblings (from different conspecific mothers or even species) that face IIC.

Acknowledgments

We thank Arodi Prado for helping to collect the parasitoids in field; Mariana G. Zério for collaborating with the insect colonies and experiments; Alyssa De La Rosa for reviewing the English; and an anonymous reviewer for providing insightful comments and fruitful suggestions to this manuscript. This research was partially funded by “Fundação de Amparo à Pesquisa do Estado de São Paulo - FAPESP”, grant number 2007/07054-8.