SCIENTIFIC CORRESPONDENCE

starvation period, which corresponded Making a meal of mother with the weight Jost by their mothers (8.0 ± 1.7 mg, n = 16). The weight lost by SIR - All provide some form of 10 the mother during the starvation period maternal care, from weaving protective affected the duration of matriphagy and silk around the eggs to guarding and "'Cl hence juvenile survival. Lighter mothers ~ feeding spiderlings1• An unusual and a5 8 were entirely consumed earlier in the -0 extreme form of care is matriphagy, where ·5. experiment than heavier mothers, and the spiderlings consume their mother. "'Ol thus among the spiderlings of C Although widely reported1- 3, the signifi­ '> 6 these lighter, deceased mothers occurred ·~ cance of this behaviour has not been :::, earlier. Consequently, the number of explored. Here we describe a novel form "' surviving offspring at the end of the 0 4 of maternal care that includes matriphagy ci experiment was positively correlated with in the Australian social Diaea z the weight lost by the mothers over the ergandros Evans (Thomisidae ). 2 starvation period (Fig. 2). Mothers store food in unviable trophic 0 5 10 15 20 Extreme forms of , such eggs that are never laid. The nutrients as matriphagy, may be most likely stored in these eggs appear to be convert­ Total maternal weight loss (mg) among spiders that typically produce ed to haemolymph which is consumed by FIG. 2 The number of juveniles surviving the single clutches. Additionally, matriphagy 'nursing' spiderlings. Later, the partially starvation period was positively correlated may facilitate the of social depleted eggs are consumed directly with the weight lost by mothers (r = 0 .65, behaviour9 by reducing cannibalism when the offspring eat their mother's F1,15 = 9.60, P <0 .01, n = 16). but not total among groups of siblings10·n, as occurs entire body. Mothers that store more spiderling weight (r = 0 .281, F1,15 = 1.119; in D. ergandros7 and other social food support longer periods of matri­ not significant). species6•11 • phagal feeding, thereby reducing sibling Theodore A. Evans cannibalism. oped ovaries were distorted and appeared Elycia J. Wallis Maternal care is more extensive among to be nonviable; yolky granules filled the Mark A. Elgar social than other spiders4- 6• D. ergandros cytoplasm and no nuclear material had Department of Zoology, mothers build a large, protective nest been stained. University of Melbourne, with eucalyptus leaves, and capture large Spiderlings do not obtain the nutrients Parkville, Victoria 3052, prey to feed their brood. The contained in these oocytes by simply eat­ Australia survival of spiderling D. ergandros ing them because, unlike trophic eggs9, 1. Foelix, R.F. Biology of Spiders (Harvard Univ. Press, depends on the mother's presence until the oocytes are larger than the oviducts Cambridge, 1982) they reach their fourth instar, after which and cannot be laid. Instead, the fourth­ 2. Barnes J. The Complete Works of Aristotle (Princeton instar spiderlings acquire these nutrients Univ. Press, 1984) they are carable of foraging and nest 3. Bristowe, W. S. The World of Spiders (Collins, London, construction ·8 • when they slowly cannibalize their 1958). We examined the ovaries of female D. mother. Initially, they imbibe small 4 . D'Andrea, M. Monitore zoo/. ital. NS Monogr. 3 (1987). 5 . Buskirk, R. E. in Social (ed. Hermann, H. R.) ergandros at different reproductive stages quantities of haemolymph from the leg 281-387 (Academic, London. 1981). (Fig. 1). Gravid females had elongated joints of the living, unresisting mother. 6. Seibt, U. & Wickler, W. Anim. Behav. 35, 1903-1905 (1987). ovaries containing many small oocytes. As matriphagy continues, the mother 7. Evans, T. A. Rec. W. Aust. Mus. (Suppl.) 51, 151- 158 Female D. ergandros lay only a single loses mobility and her opisthosoma (1995). 8 . Main, B. Y. in Australian Arachno/ogy (eds Aust in, A. D. & clutch, unlike other thomisids3, and after shrinks, presumably as the trophic eggs Heather, N. W. ) 55-73 (Aust. Ent. Soc., Brisbane, oviposition their ovaries become shriv­ are transformed into haemolymph. After 1988). elled, with no oocytes. However, mothers several weeks, she is decrepit, unable 9. Crespi, B. J. in Cannibalism: Ecology and Evolution among Diverse Taxa (eds Elgar, M. A. & Crespi, B. J.) regain weight after oviposition and have to move, and the offspring eat her 176- 213 (Oxford Univ. Press, 1992). unusual ovaries that contain significantly entirely. 10. Alexander, R. D. A. Rev. ecol. Syst. 5 , 325-383 (1974). 11. Elgar, M. A. & Crespi, B. J. (eds) in Cannibalism: Ecology fewer (t = 4.22, P < 0.01) and larger (t = The growth and survival of mothers and Evolution among Diverse Taxa 1-12 (Oxford Univ. 20.15, P <0.001) ooc)'tes than those of and spiderlings was examined for 6 weeks. Press, 1992). gravid mothers. Histological examination The brood gained weight (6.5 ± 2.8 mg, revealed that the oocytes in the redevel- n = 16) during the first 3 weeks of the Opening up ca2+ stores with lnsP3

SIR - Hirose and Iino1 introduced the low-affinity Ca2+ -sensitive dye Furaptra into intracellular Ca2+ stores of vascular smooth-muscle cells and measured the kinetics of inositol trisphosphate (InsP3)-induced Ca2+ release. This novel technique of incubating intact cells with Ca2+ buffers like Furaptra/AM (ref. 1) or Fura-2/AM (refs 2- 5) to load them m the intracellular stores, followed by permeabilization of the plasma FIG. 1 a, A gravid female (44.54 ± 2.05 mg, n = 22), with the left ovary exposed, surrounded by midgut diverticula (m). Ovaries contained 41 ± 6.02 oocytes (n = 8) 0 .26 ± 0.01 mm in diame• membrane, allows fast measurements of 2 2 ter. b, A female one week after oviposition (22.50 ± 0 .54 mg, n = 22) with a completely eviscer• the luminal Ca + concentration ([Ca +]) in ated opisthosoma; ovaries (ov) extremely small, without oocytes (n = 7). c, A mother one month the presence of a constant cytosolic 2 2 after oviposition (47.13 ± 1.69 mg, n = 22) with en larged, irregular oocytes. Ovaries contained [Ca +]. As luminal Ca + can control InsP3- 13 ± 1.51 oocytes (n = 8) 0.77 ± 0.03 mm in diameter. Scale bars, 1 mm. induced Ca2+ release6, we investigated in NATURE · VO L 376 · 27 JULY 1995 299