S. Afr. J. Antarct. Res., Vol. 13, 1983 45 A note on the feeding of marioni Vietti larvae at Marion Island

D.D. French Institute of Terrestrial Ecology Banchory, Kincardineshire AB3 4BY, Scotland and V .R. Smith Institute of Environmental Sciences University of the Orange Free State, Bloemfontein 9301

Larvae of the Lepidopteran Viette form container (100 ml), together with li ving and dead P. eookii an important component of the soil macrofauna at sub· roots. green and dead P. eookii leaf bases, some peat and one Antarctic Marion Island. They were previously considered as earthworm. After one night (in the laboratory at ea 16 °C). the herbivores, feeding on plant roots and rhizomes. Captive larvae worm was almost completely (ea 80% ) consumed and all the are omnivorous and will aggressively attack and devour other foodstuffs uneaten. After the second night the wonn had earthworms. The significance of this carnivory in the field is disappeared , the roots had been eaten, the dead leaves had unknown bw the larvae and earthworms occur together in the only been nibbled and the green leaves were completely upper peat layers ofm any of the island plant communities. untouched. A ll remaining items were then completely ignored by the larvae until they died on the fourth day. Larwes van Pringleophaga marioni () vorm 'n A further trial, in which the larvae were subjected to temp­ belangrike komponent van die grond-makrofauna van die sub­ erature and humidities closer to those experienced in the field, Antarktiese Marion-eiland. Die larwes is vantevore as herbi· was also performed. One larva was placed in each of three vore war plantworrels en worrelsrokke vreet, beskou. Larwes separate containers, each with: wat eksperimenteel afgesonder word, is egter omnivore en sal 2 cm green Poa leaf base selfs erdwurms aanva/ en verorber. Die is nie bekend of 2 cm dead Poa leaf base erdwurms in die naruur deur die larwes benut word nie. maar ea 1 cmJ Clasrnatocolea humilis (common bryophyte in mire lan ves en erdwurms kom saam in die hoonste veenlae van baie areas) 3 van die eiland se plantgemeenskappe voor. ea I cm peat 2 small earthworms (ea 4 cm, probably Mieroseolex kerguelarum Grube). Introduction The containers were placed in an incubator at 10 oc and 75-85 Larvae of the flightless Lepidopteran Pringleophaga marioni per cent relative humidity. The initial 15 hours of this trial Yiette are an important element in the macroinvertebratc were conducted in the dark and the remainder in subdued fauna of Marion Island , being one of the most abundant light. At irregular intenals throughout the trial the beh3viour groups after earthworms. They are a significant item in the of the was noted. diets of sheathbills (Burger 1978) and house mice (Gleeson Results and discussion 1981) at the island. T he life history and feeding ecology of P. marioni arc largely unknown. Van Z inderen Bakkcr (1966) The results arc presented in Table I. They show that observed the larvae among grass roots and on moss cushions Pringleophaga larvae will cat earthworms if available. Two but provided no details of what they were eating. Viettc ( 1948) larvae preferred worms over all the other foods presented suggested that the main foods of larvae of P. kergue/ensis, with while the third gave worms high priority as second choice after which P. marioni may be conspccific. arc the roots and roots. Feeding only commenced 4- 15 hours after the start of rhizomes of Pringlea antiscorbwiea and Acaena magellaniea. the trial. Since caterpillars generally feed continuously they According to Huntley (i 971) Pringle.ophaga marioni adults may have been driven through starvation to consume some of may be poll inators of Prinf!,fea; thus the association of the the food items presented. an acknowledged criticism of larvae with this plant species may be coincidental. ··cafetaria"-type feeding experiments. Although two larvae While dissecting tussocks of Poa cookii on Marion Island in started feeding on live roots. in only one case did this result in May 19R2, a Pringleoplwga larva was observed amongst the a l<.trgc proportion of the root material being consumed. roots, apparently feeding on an earthworm. This larva plus Average times to first noticeable consumption were : several others were placed in a plastic container together with earthworms 17 hours. live roots 30 hours. dead roots 47 hours. a living earthworm . Three of the larvae promptly converged leaves not consumed. Using thi~ measure. earthworms are upon the worm and devoured it. approximately twice as attractive as food as are roots. The li vcn~' Ort was not consumed hut no conclusions regarding the Experimentation peat can be made. Certainly the larvae were never actually seen to cat peat during the obscrv

Table I Behaviour of P. marioni larvae and consumption of food items during a 60 hour feeding trial. Time Larva 1 Larva 2 Larva3 Start Exploratory movements, showed As! As! interest in worms, roots and Clasmatoco/ea. 4hours No change. No change. No change. 15 hours No change. Live root mostly eaten. Live root slightly nibbled. Both 2 + 3 making sporadic bites at the earthworms which quickly withdraw. Clasmatoco/ea and peat removed since the larvae were apparently showing no interest in them and the peat was being spread around the container. 16hours ''Hunting" movements ini ti ated, then Aggressively biting at worm, much As2 reared up and pounced on a worm, more frequently than before. almost cutting it in two. 24hours One worm ¥.! eaten. One worm visibly bi tten in several One worm in two mangled pieces. places. 40hours One worm disappeared, the other Both roots gone. Both worms still Only small fragments remaining of bisected. Roots and leaves not present. both worms. recognisably touched. 60hours Second worm approximately Y! eaten. One worm in pieces. second has Remains of worms liquified. No other No other change except living and the tail end missing. change. dead roots nibbled.

a worm , would bite at it. The violent reaction of the worm larvae of some case-bearing feed on fibres. Since would cause the larvae to retreat. This continued until the some Lepidopterans are cannibalistic under stress, this unex­ larvae happened to bite deeply into the worm, when they pected carnivory could have been aberrant behaviour fo llow­ would immediately become aggressive and, despite the ing the disturbance of their removal from the grass tussock. It worm's struggles, tear at it in a truly ferocious manner. Before is strongly urged that the "cafetaria" trials reported on here be this stage, however, any resistance on the part of the worm was extended, as part of the recently initiated Marion Island sufficient to discourage the larvae. The worm will , therefore, entomological programme, to incorporate a wider range of more easily escape attack by Pringleophaga larvae in the field; food items (particularly bryophytes) and single-item survival hence they may not form as important a component in the tests. as well as microscopic examination of the gut contents of larval diet as suggested from the feeding trials. P. marioni larvae collected from a range of habitats. Pringleophaga larvae and earthworms are very common in the upper peat layers of many vegetation types at Marion References Island, and both come to the surface amongst the bases of Poa BURGER, A.E. 1978. Terrestrial invertebrates: a food source for cookii tussocks. The larvae have generally been thought to be birds at Marion Island . S. Afr. J. Antarct. Res. 8: 87-99. herbivorous on a limited number of sub-Antarctic plant GLEESON , J.P. 1981. The ecology of the house mouse. Mus species, but, because of their widespread occurrence over a musculus Linnaeus, on Marion Island. Unpublished M.Sc. thesis, range of vegetation types, they may be catholic detritivores University of Pretoria, Pretoria, South Africa. rather than stri ct herbivores. If so, then the step to facultative HUNTLEY , B.J . 1971. Vegetation. In: Marion and Prince Edward predation on the abundant earthworms might not be a large Islands; report on the South African biological and geological one. One observation and two "cafetaria" trials with li mited expedition 1965-1966, eds E.M. van Zinderen Bakkcr, J.M. Winterbottom and R.A. Dyer. A. A. Balkema, Cape Town, p. 98- replication are obviously inadequate upon which to base any 160. conclusion. However, the consistency of the results suggests VAN ZINDEREN BAKKER, E.M. Sr. 1966. Die nora en fauna van that Pringleophaga marioni may be an important carnivore in die sub-Antarktiese eilande Marion en Prins Edward. Tegnikon the trophic structure of the Marion Island terrestrial eco­ 15: Ll7. system. To the authors' knowledge this is the first observation VIETIE, P. 1948. Croisierc du Bougainville aux iles australes fran· of carnivory in Pringleophaga, and they are also not aware of caises. XX. Lcpidoptere. Mem. Mus. Nam. Hist. nat., Paris (N.S.) any record of preying upon other , although 27: 1-28.