Behavioral Ecology and Sociobiology

, Volume 56, Issue 3, pp 201–206

Nest-size variation reflecting anti-predator strategies in social spider mites of Stigmaeopsis (Acari: Tetranychidae)

Authors

    • Laboratory of Animal Ecology, Graduate School of AgricultureHokkaido University
    • Symbiotic Engineering, Department of Bioinformatic Engineering, Graduate School of Information Science and TechnologyOsaka University
  • Yutaka Saito
    • Laboratory of Animal Ecology, Graduate School of AgricultureHokkaido University
Original Article

DOI: 10.1007/s00265-004-0776-7

Cite this article as:
Mori, K. & Saito, Y. Behav Ecol Sociobiol (2004) 56: 201. doi:10.1007/s00265-004-0776-7

Abstract

The social spider mites (Acari: Tetranychidae) of Stigmaeopsis weave dense nests on the underside of host leaves. Four species occur on the leaves of bamboo in Japan: Stigmaeopsis longus, S. celarius, S. takahashii and S. saharai. We initially reconfirmed the occurrence of distinct variation in nest size among the species. Based on the hypothesis that this variation plays a role in protecting the spider mites from predators, we looked at the behavior of the natural enemies that occur on the host plants along with members of Stigmaeopsis. We found considerable variation in the ability of nests to protect the spider-mite eggs. The smallest nests protected the eggs against three predators, whereas the largest nests protected the eggs against only one predator species. So, decreases in nest size increased egg defense. Thus we concluded that nest-size variation reflects a strategy for reducing predation.

Keywords

Spider mitesSilken web nestProtective refugePredator-prey interactionStigmaeopsis

Introduction

In animals other than those at the highest trophic level, defense mechanisms are one of the most important adaptations for survival, and extraordinary variation in the means of defense has been reported in arthropods (Evans and Schmidt 1990), as well as in other animals (Alcock 1989). In a review of lepidopteran insect defense, Lederhouse (1990) said: “Although many of the more spectacular adaptations have engendered ‘just-so’ adaptive stories, experimental demonstrations of their efficacy are limited. Well-designed experimental studies and perceptive field observations are required if we hope to make substantial progress in understanding the role of diverse primary defense mechanisms in reducing predation.”

Most social animals inhabit fully or semi-permanent nests (e.g. Seger 1991). Since the nest structure determines the social unit in which individuals will interact with each other, the forces molding nest morphology should influence social behavior and even social organization. In plant-parasitic arthropods, many aphids, thrips and mites form nest-like structures and include several social species (Saito 1986a; Crespi 1992; Foster and Northcott 1994; Crespi and Mound 1997; Saito 1997; Stern and Foster 1997). The nest (or gall) that is constructed by such animals is thought to have at least two important functions: feeding site and protective refuge (Dixon 1973; Saito 1985; Price et al. 1987; Crespi 1994; Foster and Northcott 1994). The diversity of nest (gall) forms in such animals is supposed to be maintained primarily through natural selection for such functions (Price et al. 1987). The functions of protective refuges are, however, not always clear because it is usually difficult for nest-making animals to live normally under “non” nest conditions, i.e. it is often difficult to settle on appropriate controls for the study of nest function. In fact, several experiments, which could have revealed the functions of traits for protecting young, were limited to studies that used artificial “trait-removal” experiments (Kudo and Ishibashi 1996; Mori et al. 1999; Toyama 1999; Yanagida et al. 2001). If there is great variation in nest characteristics such as size and structure among very closely related and sympatric species, a comparison of the shared ecologies should make it possible to detect the function of the nest.

The nest-weaving spider mites of Stigmaeopsis (Saito et al. 2004) weave dense nests over depressions on the underside of host leaves (Fig. 1). This genus, which infests dwarf bamboo and bamboo plants, can be classified into four species: Stigmaeopsis longus, S. celarius, S. takahashii and S. saharai according to nest size (Saito and Takahashi 1980; Takahashi 1987). In addition, a molecular phylogeny study using partial sequences of ribosomal DNA (945 bps) demonstrated that these species are close sibling species in Tetranychidae (Sakagami 2002). Moreover, reciprocal cross experiments supported the results of phylogenetic analysis because reproductive isolation between S. takahashii and S. saharai was incomplete (Mori 2000). Therefore, we thought that a comparison of the effects of different nest sizes among these species could demonstrate the function of nests.
Fig. 1

Nesting patterns of Stigmaeopsis

Saito (1986a, 1986b) reported that there is biparental defense against the predatory mite, Typhlodromus bambusae, in the social spider mite, S. longus, which has the largest nest size among these species. This counterattack behavior is known to be effective if many individuals cohabit in the same nest, and thus the group living guaranteed by a large nest is supposed to be adaptive. Furthermore, Mori et al. (1999) demonstrated by a series of removal experiments in a natural forest that both the web and female attendance of S. longus improve the survival of young. In this context, we had doubts as to why the other three species, which make smaller nests, sometimes cohabit with S. longus on the same dwarf bamboo leaves, as it shows no counterattack behavior (Mori 2000). In order to answer this question, we attempted to determine how the nest size differences known in the four species are effective for protection against predatory intrusion under experimental conditions.

Methods

Biological remarks on the species

S. longus occurs on the leaves of the dwarf bamboo plants, Sasa senanensis and Sasa kurilensis in Hokkaido, Japan and makes very large nests (Saito and Takahashi 1980; Saito 1986a). There are three other species: S. celarius, S. takahashii and S. saharai. S. celarius mainly occurs on Moso bamboo, Phyllostachys pubescens, and Machiku bamboo, Dendrocalamus latiflorus, in southern Japan, although it can develop and reproduce normally on S. senanensis and S. kurilensis (Mori 2000). S. takahashii usually occurs on S. senanensis. S. saharai mainly occurs on S. senanensis and S. kurilensis in Hokkaido, Japan and makes the smallest nests of Stigmaeopsis. The nest sizes of these species are in the order of S. longus>S. celarius>S. takahashii>S. saharai (cf. Fig. 2). The morphological characteristics and life histories of all species are very similar except for the lengths of certain dorsal setae (Saito and Takahashi 1982; Saito 1990; Mori 2000). The length of dorsal seta P2 shows four separate frequency distributions according to the species and this seta length is correlated with nest size (Saito and Takahashi 1980). Group size (i.e., number of individuals per nest) is basically proportional to nest size, and it can often become very large in S. longus and S. celarius through nest-extension (“united nest” in Fig. 1). S. longus or S. celarius individuals live gregariously inside web-nests throughout their lifetimes and show cooperative nest-defense behavior against predators (reported in the former by Saito 1986a and in the latter by Mori 2000). However, because the nests of S. takahashii and S. saharai are smaller than those of S. longus and S. celarius, S. takahashii and S. saharai females repeatedly disperse and found new nests during their lifetimes (Saito and Takahashi 1982). These facts indicate that the variation among species of Stigmaeopsis in the characters mentioned can be accounted for by the selection pressures that cause a change in nest size.
Fig. 2

Mean areas (±SE) per nest of Stigmaeopsis. The results of post-hoc tests (Scheffe’s method) are shown (n=numbers of females analyzed)

The predatory mites used in this study all inhabit S. senanensis and they were often observed with the mites of Stigmaeopsis. At least seven spider mite species and seven predacious mite species (including five species used in this study) occur on S. senanensis.

Materials

All spider-mite species used in this study were the progeny of those obtained from field populations, as shown in Table 1. These mites were reared in the laboratory under conditions of 23±2°C, 40–80% R.H. and 15L-9D. The host plants for the cultures and experiments (S. senanensis and S. kurilensis) were cultivated in a greenhouse.
Table 1

Collection and culture record of Stigmaeopsis used in this study

Culture ID

Species

Date

Locality (Prefecture)

Host plant

Plant for culture

A

Stigmaeopsis longus

May 10 1997

Sapporo (Hokkaido)

Sasa senanensis

Sasa senanensis

B

Stigmaeopsis celarius

March 30 1998

Toyonaka (Osaka)

Phyllostachys pubescens

Sasa senanensis

C

Stigmaeopsis celarius

March 30 1998

Toyonaka (Osaka)

Phyllostachys pubescens

Sasa kurilensis

D

Stigmaeopsis takahashii

May 12 1997

Sapporo (Hokkaido)

Sasa senanensis

Sasa senanensis

E

Stigmaeopsis takahashii

June 15 1998

Sapporo (Hokkaido)

Sasa senanensis

Sasa senanensis

F

Stigmaeopsis saharai

June 9 1998

Higashikawa (Hokkaido)

Sasa senanensis

Sasa senanensis

The predatory mites used in the experiments were collected on S. senanensis and P. pubescens which most of the spider mites used in this experiment inhabited sympatrically (Table 2). They were reared on the detached leaves of S. senanensis using the four species, as well as another co-occurring species, Yezonychus sapporensis, for prey.
Table 2

Collection record of predatory mites used in this study

Species

Date

Locality (Prefecture)

Host plant

Typhlodromus bambusae

May 12 1997

Sapporo (Hokkaido)

Sasa senanensis

Typhlodromus bambusae

March 30 1998

Toyonaka (Osaka)

Phyllostachys pubescens

Agistemus iburiensis

September 4 1997

Sapporo (Hokkaido)

Sasa senanensis

Phytoseius tenuiformis

May 12 1997

Sapporo (Hokkaido)

Sasa senanensis

Agistemus summersi

May 12 1997

Sapporo (Hokkaido)

Sasa senanensis

Amblyseius sp.1

August 4 1997

Sapporo (Hokkaido)

Sasa senanensis

Evaluation of nest size

The prey populations of A, C, E, and F in Table 1 were used. The experiments were conducted under conditions of 25±1°C, 50–70% R.H. and 15L-9D. Adult females aged 1 day subsequent to mating (i.e., just after the final molt) were taken from stock cultures of each species and placed individually onto detached leaves (ca. 1.5 cm2) of host plants (S. senanensis or S. kurilensis) placed on water-soaked cotton sheets which had been spread on polyurethane mats in petri dishes. In order to prevent water evaporation from lowering the temperature of the leaf surface (Saito and Suzuki 1987), the surfaces of the cotton and polyurethane mats surrounding the detached leaves were covered with polyethylene film. After 48 h, we measured surface area (=maximum length×maximum width) of each nest with a divider, because the nest size of S. longus is known to be stable ca 24 h after commencement of nest building (Y. Saito, unpublished data; K. Mori, unpublished data).

Protection efficiency of nest

The prey populations of A, C, D, and F in Table 1 were used for this experiment. A detached leaf measuring 3 cm×3 cm was prepared by surrounding it with water-soaked cotton. Seven to 10 females of each species were introduced from stock cultures onto experimental arenas and kept under conditions of 23±1°C, 40–80% R.H. and 15L-9D for 3–4 days. After the females had constructed their nests and deposited a sufficient number of eggs within them, we removed the females from the leaves and prepared two treatments, i.e. the “web-removal” and “web-intact” treatments.

Gravid female predatory mites of each species (Table 2) were introduced individually onto each experimental arena. Forty-eight hours after predator introduction, we recorded the number of prey eggs eaten, the number of eggs laid by the predator and the location of the predator female and her eggs (inside or outside the web-nest in the “web-intact” treatments). These parameters are the criteria for predator intrusion. In particular, the number of eggs consumed indicates the intensity of predation of the respective spider mites.

Statistical analyses were carried out using the software, StatView for Windows (Abacus Concepts, Berkeley, Calif.).

Results

Nest-size variation

The nest sizes constructed by females of each species are shown in Fig. 2. There was a significant difference in mean nest area (Fig. 2) among the four species (one-way ANOVA, F3, 122=144.84, P<0.0001). Multiple-comparison tests (Scheffe’s method) detected significant differences between all combinations other than between S. celarius and S. takahashii (Fig. 2).

Effect of webs

When no web was present, the eggs of all four species were fed on by all predators (Fig. 3), indicating that each predator species can eat the eggs of each spider mite without difficulty. All the predatory species used in this study are potential natural enemies for all prey species. Several predator species showed preferences for certain prey-species eggs: more S. longus eggs than those of S. celarius and S. takahashii were eaten by Phytoseius tenuiformis (one-way ANOVA, F3, 60=4.50, P=0.0065; Scheffe’s test, S. longus vs S. celarius: P=0.036, S. longus vs S. takahashii: P=0.024). More S. longus eggs than S. celarius and S. saharai eggs (one-way ANOVA, F3, 61=6.39, P=0.0008; Scheffe’s test, S. longus vs S. celarius: P=0.005; S. longus vs S. saharai: P=0.0374), and more S. takahashii eggs than S. celarius eggs were eaten by Agistemus summersi (Scheffe’s test, S. takahashii vs S. celarius: P=0.0373). The other predators consumed the eggs of all prey species indiscriminately.
Fig. 3

Mean numbers (+SE) of four species’ eggs eaten by five predators in “web-removal” treatments. Numerals over columns are the numbers of replicates

In the “web-intact” experiments, it was shown that the nest of each spider mite greatly affected predation efficiency (Fig. 4). In particular, Amblyseius sp. 1, which was the most serious predator for all species in the “web-removal” experiments, could not consume any prey eggs. The other predator species could feed on the prey eggs inside nests. We then applied a two-way ANOVA to compare the number of eggs eaten during 48 h by each predator between “web-removal” and “web-intact” treatments, and between prey species. A. iburiensis females consumed prey eggs regardless of web presence (“presence of web” effect: F1, 133=3.43, P=0.066). T. bambusae females also consumed prey eggs regardless of web presence (“presence of web” effect: F1, 128=0.72, P=0.397) while the interaction between the effects of “prey species” and “presence of web” was significant (F3, 128=11.98, P<0.0001). However, P. tenuiformis and A. summersi females were blocked by prey web-nests (“presence of web” effect for P. tenuiformis: F1, 115=138.33, P<0.0001; “presence of web” effect for A. summersi: F1, 122=110.90, P<0.0001), although the interaction between effects of “prey species” and “presence of web” for A. summersi was also significant (F3, 122=2.97, P=0.034).
Fig. 4

Mean numbers (+SE) of four species’ eggs eaten by five predators in “web-intact” treatments. Numerals over columns are the numbers of replicates

In order to evaluate the overall effect of web-size variation on the protection of eggs from predators during 48h, we applied a two-way ANOVA to the data of all predators in “web-intact” treatments except Amblyseius sp. 1, which never ate eggs in web-nests. There was a great difference in the effect of web types (“prey species” effect) on the number of eggs eaten (by two-way ANOVA, “prey species” effect: F3, 250=7.97, P<0.0001, “predator species” effect: F3, 250=158.82, P<0.0001, interaction between “prey species” and “predator species”: F9, 250=24.21, P<0.0001). Paired comparisons by the use of post-hoc testing showed that there were significant differences in predation pressure between S. longus and S. celarius (Scheffe’s test, P<0.0067) and between S. longus and S. saharai (Scheffe’s test, P<0.0206). S. longus eggs were preyed upon by four predator species: (T. bambusae, A. iburiensis, P. tenuiformis and A. summersi), and those of S. celarius and S. takahashii by two predator species: (T. bambusae and A. iburiensis). However, the eggs of S. saharai were preyed upon by three predator species: (T. bambusae, A. iburiensis, A. summersi), though an A. summersi female sometimes preyed on S. saharai eggs directly through the nest roof without intruding into the S. saharai nest. These results could be summarized as decreases in nest size increased egg defense.

The proportion of predator females and their eggs observed inside the nests 48 h after their introduction in “web-intact” experiments corresponded well to the pattern of prey-egg consumption (Table 3). Most T. bambusae and all A. iburiensis females and all their eggs were observed inside the nests regardless of prey species. In addition, many P. tenuiformis and A. summersi females (and their eggs) also remained in S. longus nests, although no females from either of these two predator species were observed in the web-nests of the other species.
Table 3

Percentage of predator females (adults) and laid eggs inside nests (%). Numerals in parentheses are the numbers of females or eggs tested

Prey species

Predator stage

Predator species

T. bambusae

A. iburiensis

P. tenuiformis

A. summersi

Amblyseius sp.1

Stigmaeopsis longus

Adult

95.24 (21)

100 (22)

65 (20)

66.67 (18)

0 (13)

Egg

100 (40)

100 (131)

88.24 (34)

70.97 (62)

0 (12)

Stigmaeopsis celarius

Adult

100 (12)

93.75 (16)

0 (10)

0 (14)

0 (14)

Egg

100 (47)

100 (96)

0 (7)

0 (34)

0 (14)

Stigmaeopsis takahashii

Adult

94.74 (19)

100 (20)

0 (20)

0 (20)

0 (16)

Egg

100 (58)

100 (123)

0 (7)

0 (43)

0 (6)

Stigmaeopsis saharai

Adult

93.75 (16)

100 (16)

0 (9)

0 (13)

0 (13)

Egg

100 (62)

100 (80)

0 (3)

0 (23)

0 (6)

Discussion

We have been able to demonstrate that the nest-size variation of Stigmaeopsis showed great differences in the effectiveness of predator avoidance. First we discuss why such variation has evolved in this mite group.

If the web-nest functions as a protective refuge against predator intrusion, then variation in its structure will change the actual predation pressure from which spider mites have usually suffered in a particular environment. As known previously, the fauna of phytophagous mites on Sasa plants is very complex (at least seven spider-mite species and several eriophid species) and varies greatly depending upon the leaf-hair density of hosts (Chittenden 2002). The fauna of predacious mites also varies due to the change in prey composition. In such a situation, it is possible to hypothesize that the nest-size variants were selected under various environments. For example, they may interact with different predators and/or phytophagous animals in different environments. So, the species making smaller nests were generated under stronger predation pressure from generalist predators. In this case, the species have differentiated allopatrically.

However, some populations of S. longus live with S. takahashii or S. saharai in the same natural forests (Takahashi 1987 and Table 1). This suggests that there are alternative anti-predation strategies at work in such forests. As shown in the present study, the nests of S. takahashii or S. saharai are very effective as protective refuges against several predator species, such that these species have a great advantage over S. longus in habitats where these predators occur. In such environments, the predation pressure on S. longus must be very intense, such that it has needed to develop additional anti-predator adaptations. The counterattack behavior performed by adult males and females (biparental defense) of S. longus (Saito 1986a, 1986b) is thus thought to be another strategy for improving the survival of this species. It should be noted here that the effect of such counterattack behavior increases with the increases in mite density within a nest and with the staying time of parents in a nest, conditions that are only realized in large nests (Saito 1986b; Mori 2000). Therefore, we could conclude that there are at least two extremes in strategies, i.e., “protection by smaller nests” in S. takahashii and S. saharai and “defense by many individuals in larger nests” in S. longus and S. celarius.

This involves a very important view regarding social evolution. Spatio-temporal aggregation and biparental defense are two sets of traits that characterize the sociality of S. longus (=L form, Saito 1997). However, small nests with a small number of eggs and short maternal attendance duration in S. takahashii and S. saharai mean that there is no highly developed sociality in these species, even though they are sub-social (Saito 1995). Therefore, the alternative anti-predatory adaptations are thought to be connected to the kinds of sociality that have evolved in these mite species; namely, the anti-predation adaptations are primarily responsible for the evolution of sociality in this mite group.

So far, we have discussed only the effectiveness of the self-constructing protective refuge and sociality of four species, but there seems to be an additional point of view of anti-predator behavior in these four species, i.e. aggregation and dispersion. The function of aggregation has received considerable attention as anti-predator behavior (Hamilton 1971; Pulliam and Caraco 1984; Inman and Krebs 1987; Vulinec 1990) and two distinct mechanisms can be involved: “encounter effect” and “dilution effect”. These were combined and termed “attack abatement” by Turner and Pitcher (1986) (see also Wrona and Dixon 1991; Uetz and Hieber 1994). We have also studied the effect of egg-deposition patterns in relation to nest size under experimental conditions, with the result that smaller nests distributed sparsely (i.e., ovipositing eggs in small clumps) effectively decrease the probability of predation (deluding effect, Y. Saito et al. unpublished work).

We thus believe that the effects of counterattack (Saito 1986a, 1986b), making protective refuges (in this study) and egg-deposition patterns are all anti-predatory strategies and may explain why there is variation in nest size in Stigmaeopsis.

Lastly, we should discuss the mechanisms of how the smaller nests effectively prevent predator intrusion. The nest-size variation referred to in this study involves several differences in the nest structure. For instance, the nest size may be closely related to nest-entrance size. Furthermore, the density of the silken threads of the nest web may be higher in small nests, if the investment in nest construction is equal in all species. Although we have no quantitative data, we believe that the size of the nest entrance is primarily responsible for the effectiveness of anti predatory intrusion, because most predators usually tried to enter from the nest entrance.

Acknowledgements

We thank Drs. Y. Watanuki, E. Hasegawa, T. Nagata, T. Gotoh, T. Sakagami, and A.R. Chittenden and Ms. Y. Sato for valuable suggestions. This study was partly supported by Grant-in-Aid for Scientific Research (B), no. 13440227 from the Japan Society for the Promotion of Science. Research activities were conducted in compliance with the current laws of Japan.

Copyright information

© Springer-Verlag 2004