Mite species

To study the effects of mating status on starvation tolerance and longevity, we used the two-spotted spider mites of which the mating behaviour and reproductive system were well known. The spider mites used in the study were from a laboratory-reared population established on common bean (Phaseolus vulgaris L.) for nearly 3 years in a greenhouse of Manaaki Whenua – Landcare Research, Auckland.

Experiment preparation and rearing unit

At the commencement of this experiment, we randomly picked about 240 quiescent males and 240 quiescent females in their deutonymphal stages from the bean leaves using a fine brush and transferred them individually onto each leaf disc. The freshly emerged adults were used to establish the experiment. The rearing unit for spider mite fed ad libitum is a round leaf disc (80 mm in diameter) floating on the water filled in each well of the cell culture plate (24 well plate). The leaf disc was a little smaller than the well in size (well volume 3.4 ml, diameter 15.6 mm), so water around the leaf disc can function as a barrier that prevents mites escaping. Spider mites in starvation were raised in similar units, except that the leaf discs were replaced by a black plastic sheet so that the mites had access to water but no food.

Experiment manipulation

We fully crossed two factors, both with two levels: the mating status (mated=a male paired a female and virgin =isolated male or female without mating) with food availability (starvation and with food ad libitum), thus generating four treatments. Two days after the adults emerged, females and males were haphazardly assigned to each treatment. The survival of males and females was checked daily until death. The number of eggs produced by the females was also recorded daily. Mites were identified as not dead when they were active or could move when they were disturbed on their legs by a soft hairy brush. The experiment was conducted in a laboratory under the condition of 25 ± 1 °C, 12:12 light cycle, with relative humidity of 65 ± 5%.

The body width of the mites was also measured as an index of body size for each individual after death to examine whether body size was associated with the starvation tolerance and lifespan of the mites. The dead body of each mite was collected and mounted in Hoyer's medium on microslides, pressed flat, and heated in an oven at 50 °C for 2 weeks. The specimens were then measured in µm at 20-fold magnification using an optical microscope connected to a video camera with NIHimage software on a computer. The body width was measured as the distance between the bases of paired sce setae on the propodosoma (shown in Fig. 1), which is reliable index reflecting the body size of Tetranychidae (Sato et al. 1999). Initially forty eight replicates were set up for each treatment; owing to accidental death and missing, the final sample size ranged from 26 to 44.

Data analysis

The survival difference of the spider mites was analysed with Cox's proportional hazard model with mating treatment, food availability, and sex as fixed factors and the body width as a covariate. The possible three- and two-way interactions mating treatment × food availability × sex, mating treatment × food availability, mating treatment × sex, food availability × sex were also included in this model. The Kaplan-Meier survival analysis was used for comparisons of survival times across treatments, with Log Rank (Mantel-Cox) tests. The lifetime fecundity of females was analysed using the general linear models, with the mating treatment and food availability as two main factors with the interaction included. The dorsal width of spider mites was fitted in the general linear models with the mating treatment, food availability, sex, and all the possible interactions as explanatory variables. Correlations between longevity and body size were tested using the Pearson's product moment correlation.

The analyses were performed in R studio (R version 3.4.3). The Cox's proportional hazard model was performed using function “coxph” in R package survival (Therneau & Thomas, 2015), and general linear models were performed using the function “aov”. Data are given as mean± SE throughout the text, unless stated otherwise.

FIGURE 1.

Dorsal side of T. urticae showing the measured distance between paired setae sce marked with red double-headed arrows for a male (A) and female (B). The females are much larger than the males.

Effects of mating status, sex on survival time of starved mites

The longevity of mites varied significantly, with main factors including their mating status, sex, and food availability. The spider mites showed a sex difference in response to food deprivation, with the females more resilient to starvation—showing a much higher survival rate than the males during starvation (χ²= 71.5, df= 1, P <0.000, Fig. 2A) regardless of their mating status. When they were deprived of food for 2 days, the survival rates for females and males were 100% and 86.96%, respectively. The median survival time of males was 3 days on average, much shorter than females (6 days). Maximum survival time for males and females were 7 days and 11 days, respectively. There is also a significant interaction between sex and mating status on the survival of starved mites (z = 3.288, P =0.001): the female lifespan under different mating status did not differ (P >0.05), whereas the mated males showed a lower survival than the virgins (χ²= 15.4, df= 1, P <0.000).

Body size and its relationship with the survival time of starved mites

The female spider mites were significantly larger than males, with the female being about 480 µm and male being 325 µm in dorsal length (Fig. 3A). For both sexes, the survival time in absence of food did not demonstrate significant correlation with their body size indicated by the linear regression (F_1,38= 0.5918, P=0.4465 for mated female; F_1,44= 2.411, P=0.1276 for mated male; F_1,33=3.993, P= 0.05399 for virgin female; F_1,44= 0.4127, P = 0.524 for virgin male, Table1). This is consistent with the results of Cox proportional hazards model, which argues that body size showed no effects on the survival time of mites at food deprivation (z = 0.04, P =0.96).

Effects of mating status and mating on lifespan of mites fed ad libitum

Spider mites fed ad libitum lived longer than the starved ones (z = 13.1, P <0.000). When fed, the males showed a longer mean lifespan than the females (25 days for males and 12 days for females; F_1,145= 59.06, P <0.001). The mating status had a significant influence on the survival of mites fed ad libitum: mated females showed a significant decrease in longevity, while the male longevity was not significantly influenced (P>0.05)—consistent with the results obtained from survival analysis for mites fed ad libitum (Fig. 2B).

FIGURE 2.

The proportion of surviving adult mites on starvation (A) and fed ad libitum (B), with the solid red lines for the mated females, dash red line for virgin females, the solid blue line for mated males, dash blue line for the virgin males.

Body size and its relationship with the lifespan of mites fed ad libitum

As with those mites on starvation, the females were larger than the males in dorsal width (F_1,158= 740.98, P <0.001 Fig. 3B). In comparison with starved mites, the body size of females significantly increased when they were fed at adult stage (F_1,158= 40.18, P<0.001). However, food availability did not alter male body size (P>0.05). The lifespan of mites showed no significant relationship with body size within each treatment (All P>0.05, Table1), suggesting that body size did not contribute to the longevity difference of mites. Across treatments, the influence of body size on the survival of mites was also insignificant (z = -1.258, P=0.21).

Effects of food and mating status and body size on female fecundity

The starved females mostly ceased reproduction on the second day of starvation. Thus, their lifetime fecundity averaged 1.8 eggs per female, significantly lower than that of females fed ad libitum (76 eggs per female for fed female; F_1,154=117.51, P<0.001, Fig. 4). While mating status also influenced female fecundity, as the mated females produced fewer eggs than virgins (F_1,154=10.33, P =0.002), its interaction with food availability is non-significant (F_1,154=2.971, P=0.086). Regression analysis showed that for mated starved females, the relationship between lifetime fecundity and body size was not significant (F_1,38=0.60, P= 0.44, Table 2). However, the body size of mated females fed ad libitum positively correlated with their fecundity; however, the significant level was marginal (F_1,41=3.573, P=0.067). For the virgin females, there is a significant positive relationship between these two parameters, regardless of food availability (F_1,33=10.44, P=0.003 for starved virgin females; F_1,42=4,818, P= 0.033 for fed ad libitum virgin females).

FIGURE 3.

Body size (Means±95% percentile) of the mated and virgin mites under different conditions (A: on starvation; B: fed ad libitum). In both figures, the red boxes denote the mated mites, and the blue boxes represent the virgin mites.

FIGURE 4.

Fecundity (Mean±SE) of mated and virgin females as a function of food availability (feed ad libitum and starvation). Note there is a break on the y-axis, with the lower half for the fecundity of mites on starvation, and upper for the fecundity of females feed ad libitum.

TABLE 1.

The correlation between body size and adult lifespan of mated and virgin mites with different food availability (on starvation and fed ad libitum).

TABLE 2.

The correlation between body size and female fecundity of mated and virgin mites with different food availability (on starvation and fed ad libitum).

Sex-specific tolerance to starvation

In the short-term test without any food replenishment, we demonstrated that there was a remarkable sex difference in starvation tolerance: female spider mites could withstand much longer starvation than males could and had a much higher survival rate. Similar results were also shown in a copepod (Acartia tonsa), whose males had a lower survival than females in the absence of food (Finiguerra et al. 2013). This sex-specific starvation response can be attributed to their difference in energy reservation. In a previous study with fruit fly D. melanogaster, it was shown that the content of body lipid and starvation tolerance are positively associated with each other, and female fruit flies had a higher amount of body lipid than the males did (Chippindale et al. 1996), which is consistent with the size-efficient hypothesis (Threlkeld 1976; Tessier et al. 1983). This sex difference in energy storage may also occur in mites as well as copepods and thus contribute to the sex-specific starvation tolerance.

The females under starvation might be able to shift energy between soma maintenance and reproduction, as indicated by two main pieces of evidence. On the one hand, females mite significantly reduced their reproductive rates on the first day of starvation. When they fed, they were supposed to produce 6∼7 eggs on average (Li & Zhang unpublished data), but in this study the starved female laid only 1.3 eggs averagely. It is possible that starved females retain or absorb the egg when they encounter tight energy budget so that they can re-allocate the resource from reproduction to survival, just as in other arthropods (Montserrat & Magalhães 2007; Boivin & Ellers 2016). On the other hand, the females eased their reproduction on the second day of starvation, suggesting that they shifted energy for egg production to their own survival under adverse conditions.

Another possible explanation for the sex difference in response to starvation is their behavioural difference in the adult stage. Male mites are generally active in searching for females in order to sire more offspring, while female mites are less active and spend most of the time resting. This was supported by the evidence that many more male than female mites were trapped in water. In addition, the mated males showed lower survival rates than virgins. The higher energy expenditure results in an early onset of energy exhaustion and eventually death.

Body size and longevity

The body size of T. urticae varied substantially with food availability, and those females that were fed were much larger than starved females. However, the difference in body size has no clear association with their longevity and starvation tolerance, suggesting that body size is not a critical contributor to the starvation resistance and longevity of spider mite under our lab conditions. This result agrees with a number of studies that used invertebrates as models to explore size-longevity relationship at within-species level, and which found body size or mass is not a reliable/accurate predictor of survival. For example, in a recent study with female Drosophila melanogaster, longevity is not significantly affected by body size (Travers et al. 2015). The non-significant relationship between longevity and body size was also frequently reported in other model organisms such as C. elegans and non-model organisms, including the beetle Callosobruchus maculatus (for female only Fox et al. 2004), parasitoid Cephalonomia stephanoderis Betrem (Lauzière 2014), stinkbug predator Podisus rostralis (Zanuncio et al. 2002), mosquito Aedos triseriatus (Landry et al. 1988), and goldspotted oak borer Agrilus auroguttatus (Lopez & Hoddle 2014). This report may suggest that within species the morphology variation in body size is not adequate to result in lifespan changes. The interspecific positive relationship between body size and longevity was based on the oxidative stress theory of ageing. It states that small animals have a relatively higher mass-specific metabolic rate than large species, giving rise to severer oxidative damage and shorter longevity. However, because metabolic rate is not in proportion to body size within species, there is no reason to expect that oxidative damage should relate to body size (Khazaeli et al. 2004). It is therefore likely that within species, the range of body size and the metabolic rate variation are not the decisive factors affecting longevity.

Body size and female fecundity

The body size of the female was found to be positively associated with their fecundity, particularly for virgin females, but the mated females on starvation were an exception as they produced very few eggs even when they lived for a few days. This is not surprising, given that natural selection favours large females that have more energy and nutrient reserve for egg production, meaning they have a fecundity advantage over smaller females. The difference in the relationship across treatments may suggest that other factors such as food and mating status also affect fecundity. In this study, when mated females were deprived of food they produced very few eggs because starvation made them save energy for survival. It seems that the association between body size and fecundity collapsed because of starvation or food level. This was also supported by research with Agrilus auroguttatus, in which the females with unknown feeding history collected from the field showed that lifetime oviposition was not correlated with any body size indicators, including elytron length, elytron width, or tibia length (Lopez & Hoddle 2014); earlier field research by Branson (2008) and Gotthard et al. (2007) also failed to find positive association between body size and reproductive traits.

In insects and many other animals, high stress resistance to environmental factors can enhance survival probability. Consequently, in some studies stress resistance was deemed an indicator of organisms' great fitness (Johnson et al. 2002; Lithgow & Walker 2002; Rea et al. 2005). This notion was supported by some researchers, who claimed that selected lines of long-lived animals always have better performance when experiencing stress, including starvation and high temperature or vice versa (Zwaan et al. 1995; Norry & Loeschcke 2003; Stazione et al. 2017), although it is not always the case. Our study finds no evidence that the females highly tolerant to starvation lived longer than males. Surprisingly, the females were shown to have a relatively shorter lifespan than the starvation-sensitive males, irrespective of their mating status. The inversed association between starvation tolerance and longevity between male and female was principally due to their difference in reproductive investment across food conditions. The female showed dramatic increase in reproductive output when food was available, while the reproductive investment of the male may not vary between food treatments. This assumption, however, has not yet been justified in this study owing to the lack of an appropriate method for measuring the reproductive efforts in male spider mites. To test this assumption, we propose employing other model organisms, such as cricket (Gryllus spp), whose reproductive investment of the male can be easily checked, the reproductive efforts can be measured in calling efforts by electronic monitoring device (e.g.: Zajitschek et al. 2009, Archer et al. 2012; Rapkin et al. 2018).