Post-Feeding Larval Behaviour in the Blowfly

Available online at www.sciencedirect.com

Forensic Science International 177 (2008) 162–167 www.elsevier.com/locate/forsciint

Post-feeding larval behaviour in the blowﬂy, Calliphora vicina: Effects on post-mortem interval estimates Sophie Arnott, Bryan Turner * Department of Forensic Science and Drug Monitoring, King’s College London, Franklin-Wilkins Building, 150 Stamford Street, London SE1 9NH, United Kingdom Received 9 July 2007; received in revised form 26 September 2007; accepted 5 December 2007 Available online 19 February 2008

Abstract Using the rate of development of blowflies colonising a corpse, accumulated degree hours (ADH), or days (ADD), is an established method used by forensic entomologists to estimate the post-mortem interval (PMI). Derived from laboratory experiments, their application to field situations needs care. This study examines the effect of the post-feeding larval dispersal time on the ADH and therefore the PMI estimate. Post-feeding dispersal in blowfly larvae is typically very short in the laboratory but may extend for hours or days in the field, whilst the larvae try to find a suitable pupariation site. Increases in total ADH (to adult eclosion), due to time spent dispersing, are not simply equal to the dispersal time. The pupal period is increased by approximately 2 times the length of the dispersal period. In practice, this can introduce over-estimation errors in the PMI estimate of between 1 and 2 days if the total ADH calculations do not consider the possibility of an extended larval dispersal period. # 2007 Elsevier Ireland Ltd. All rights reserved.

Keywords: Dispersal; Accumulated degree hours; ADH; Accumulated degree days; ADD; Forensic entomology; PMI; Blowﬂies; Calliphora

1. Introduction have a longer dispersal phase. Nuorteva [1] provides an interesting link between distance and time by noting that In the blowfly lifecycle (from egg, through three feeding Lucilia sericata in Finland travels at the slow rate of about 1 m larval stages, a pupal and finally the adult stage) there is an per day over a moss covered forest floor, but in most other cases important late larval period when feeding ceases and the so there is no link available between distance and time, or called post-feeding larvae move away from the corpse on which importantly with temperature. they have been feeding to find a suitable site for pupariation. A Some studies, specifically on this post-feeding larval number of previously published forensic entomology studies dispersal phase [2–5], have emphasised the patterns and spatial have recorded details of the length of time or the distance distribution of larval movement. Whilst this is of considerable travelled during the post-feeding phase of larval blowflies. use in forensic entomology, particularly in identifying the most These are summarised in Table 1. Often precise details are probable areas to look for puparia in relation to a corpse, only lacking, as to, for instance, which species the measurements Gomes and Von Zuben [3] touch on the relevance of the relate to, and in others the account is primarily descriptive. dispersal stage to the estimation of the post-mortem interval What is clear is that some blowfly species (e.g. Protophormia (PMI). They show that dispersal puts an energetic cost on the terraenovae, Fannia sp. Chrysomya rufifacies and Chrysomya post-feeding larvae which causes a reduction in pupal weight albiceps, listed towards the top of the table) appear to spend less with increasing distance dispersed. time dispersing since they remain close to or on the larval food One common, but by no means foolproof method (see for source, whilst others (eg. Lucilia and Calliphora spp.) appear to example [6]) used by forensic entomologists to estimate the PMI, is the calculation of the accumulated degree hours (ADH) necessary to reach a specific point in the blowfly’s development * Corresponding author. (see Higley and Haskell in [7]). Such baseline data are normally E-mail address: [email protected] (B. Turner). established in the laboratory where, in controlled conditions, it

Table 1 Published information on the length of time or distance travelled in the post-feeding dispersal phase in larval blowflies Taxon Time Location/distance Ref. Protophormia terraenovae On corpse [14] Fannia sp. On corpse [1] Chrysomya rufifacies In clothing on corpse [7] Calliphorinae – ‘Under or near’ [15] Chrysomya rufifacies ‘Under or near’ [16] Blowfly larvae ‘In nearby soil’ [17] Chrysomya albiceps ‘Under and around’ [18] Chrysomya albiceps <20 cm (experimental arena) [3] Cochliomyia hominivorax 0.6–2 m [19] ‘Prepuparial Diptera’ 3 m [16] Cochliomyia macellaria, Lucilia sericata >4.6 m [4] and Phormia regina Calliphorinae >6m(>20 ft) [20] Lucilia sericata 6.4 m over soil [21] Calliphoridae and Sarcophagidae ‘‘Normally pupation occurs in soil but in domestic situations suitable sites may be difficult to locate [22] and fully grown maggots may be found wandering in quite unlikely situations some distance from their larval food’’ Lucilia sericata and Calliphora vicina 3–8.1 m [23] Chrysomya rufifacies <3.3 m [23] Cochliomya macellaria <5.1 m [23] Chrysomya megacephala 7 m at one site and approx. 25 m at another on lava [16] ‘Blowflies’ >30 m over hard ground [8] Lucilia illustris and L. caesar 3 days 3 m [1] Lucilia sericata 3–4 days (depending on temperature) [24] ‘Blowflies’ 24–72 h in culture [7] Calliphora vicina 5–14 days [11] Calliphorids Up to 4 days Some climbed 1–2 m to litter in tree forks [25] is usual that a suitable pupariation medium (sawdust, soiless 2. Methods compost or similar) is readily available and adjacent to the feeding larvae. The time spent in dispersal is therefore normally 2.1. General very short in laboratory settings and, in the overall calculation Wild C. vicina adults were trapped at several sites in London using bottle of ADH, usually ignored. traps (see [9], for the design) baited with pig’s liver and 30% sodium sulphide This contrasts with situations in the field at least for some solution. These blowflies were cultured in cages and provided with sugar and species. Post-feeding larvae may stay on the corpse or find water ad libitum. Egg-laying was induced by providing liquid liver exudate for suitable pupariation substrates very close to the corpse (for several days followed by solid pig’s liver as an egg laying stimulant. Larvae example if the corpse is in an area of well drained friable soil) or were then grown on pig’s liver in small chambers at 20 8C with an excess of food to avoid competition effects [10], to provide a source of post-feeding larvae for they may have to travel many metres if the ground is hard and the experiments. unyielding, as has been observed in the post-feeding disperal of The experiments made use of the lid of a large plastic box as the arena larvae from a corpse at the Anthropological Facility in (Fig. 1). Tennessee (Amoret Whitaker, personal communication) and The lid has a large peripheral indented groove, which normally fits the rolled noted by Green [8] Thus, in situations where pupariation is top of the plastic box on which it fits. When the lid is placed upside down the groove forms a continuous track 25 mm wide by 37 mm deep. One circuit round delayed until a suitable site has been found, there can be marked the lid is exactly 2 m although the number of circuits made by larvae was not differences in ADH estimates based on pupal and adult stages recorded. Larvae introduced into the track continued to travel around it and did between field and laboratory material. During long post- not attempt to scale the track walls. The arena was placed on the laboratory feeding dispersal periods time passes (and therefore degree- bench. Laboratory temperature was 21 Æ 1.0 8C. Humidity was not controlled hours accumulate) and energy is used up, leading to a reduction for, but was held by the building’s air conditioning system at approximately 50%rh. Following the specified time in the arena the larvae were returned to a in resources available for the pupa and adult stages. constant 20 8C and given moist peat to pupate in. This paper focuses on the impact of differing dispersal periods for larvae (both at the normal post-feeding time and 2.1.1. Experiment 1 also at an earlier stage of growth) on the further development This experiment explored the effects of differing lengths of time spent in with particular reference to the levels of error that such post-feeding dispersal on the time to eclosion and the size of the adults were situations might create in PMI estimates using ADH methods. examined. Larvae were collected from the cultures as they were leaving the food The experimental blowfly species used in this study was the source. These post-feeding larvae were ‘run’ in the trackway in small groups for 1, 2, 4, 6, 8, or 24 h before being placed in plastic boxes containing damp soil- blue bottle Calliphora vicina Robinseau-Desvoidy, a forensi- less compost as a pupariation medium and returned to the incubators to cally important and widespread species of urban areas in complete development. Control individuals (0 h disperal time) were placed temperate regions. immediately on the compost. The numbers for each experiment, were governed 164 S. Arnott, B. Turner / Forensic Science International 177 (2008) 162–167

Fig. 1. The plastic box lid arena with a close up of the channel used as the dispersal route. Dimensions are given in the text. by the numbers of fresh post-feeding larvae available from the cultures Between which at 20 8C is 460ADH. Whereas the difference in total 60 and 87 individual post-feeding larvae were ‘run’ for each time period. ADH to eclosion, seen in Fig. 2 (9777ADH for a dispersal The ADH to eclosion was recorded for all flies in the experiment. period of 1 h and 10800ADH for 24 h dispersal) between 1 and 24 h is 1023ADH, more than twice what would be expected by 2.1.2. Experiment 2 The second experiment was similar to the first but used 2nd instar feeding simple addition. larvae. This was considered as a simulation of the effects of being withdrawn Time spent dispersing as larvae clearly uses resources that from the food source through competition, rising water levels or some other would otherwise be available for pupal development and so it is occurrence which breaks the feeding cycle. Larvae for this experiment were not surprising that adult size, as measured by the d-cu cross vein collected from the cultures as 2nd instars. They were ‘run’ in the trackway for 1, in the wing (see [10], for details of method), is reduced with 2, 4, 6, 16, and 48 h. They were then returned to their food supply and followed through to adult emergence. In a further experiment 100 2nd instar larvae were longer post-feeding larval dispersal time (Fig. 3)(F = 14.62, removed from their food and left to run in the track for 96 h to see whether in the p < 0.001). absence of further food sources they could or would pupate. In experiment 2, 2nd larvae were forced to move away from their food for periods of up to 48 h and then returned to their 2.1.3. Experiment 3 food to develop as normal with no further dispersal delays. In The behaviour of larvae when dispersing, in particular whether they follow addition, 100 2nd instar larvae were left to circulate in the trails left by previous individuals, was examined using the flat central area of the plastic lid which was marked out into 4 quadrants by two lines drawn at right arena, simulating total and final removal from the food source. angles between the mid points of opposite sides. Post-feeding larvae were Only 7 survived to 96 h and these seven all pupated, having individually placed in the centre of the lid and their direction of travel traced on moulted into 3rd instars just prior to pupariation. These 7 a paper template of the arena. Where they dropped into the peripheral groove individuals have been added to the data for the 2nd instar larvae was recorded. Between each larva the plastic lid was cleaned with an alcohol starved for up to 48 h, and plotted in Fig. 4. Overall there is a wipe and left to dry. It was also randomly rotated to obviate any directional effects that might have been induced by slightly non-uniform overhead lighting. significant reduction in ADH with time spent off the food Ninety larvae were tested in this way. (F = 37.02, p < 0.001) but, using Fishers PLSD post hoc test This was then repeated but without the alcohol washing between each larva. indicates that there was no significant effect on the ADH to eclosion in the larvae with up to a 48 h delay in their 2.1.4. Experiment 4 In a second behavioural experiment two temporary aluminium foil walls were used to define a track along the long axis of the arena through the centre point. Several hundred larvae where confined to this track to create a distinct larval trail. Larvae were then individually placed in the centre of the arena and their path recorded as being on or off the trail.

3. Results

The results from experiment 1 show that time spent in post- feeding dispersal does influence the overall ADH to eclosion (Fig. 2). There is a significant tendency (F = 369.4, p < 0.001) for the total ADH estimate to be increased with longer dispersal times, although this is only apparent when dispersal times exceed 5 h (Fisher’s PLSD post hoc test gives non-significant p- values in pair-wise comparisons between the 0, 1, 2, and 4 h dispersal times). The increase in total ADH is not a simple addition of the Fig. 2. Effect of the post-feeding dispersal period on the accumulated degree hours spent dispersing. For example to take the extremes in hours to eclosion in Calliphora vicina. The error bars are Æ1.96 S.D. from the Fig. 2, between 1 and 24 h dispersal time, the difference is 23 h, mean. S. Arnott, B. Turner / Forensic Science International 177 (2008) 162–167 165

Fig. 3. Effect of time spent in larval dispersal on the final adult size, as measured by the d-cu crossvein, in Calliphora vicina. The error bars are Æ1.96 S.D. from the mean. Fig. 5. Effect of disruption to development by enforced dispersal in the 2nd instar stage on Calliphora vicina adult size. The error bars are Æ1.96 S.D. from the mean. development programme. It is only the inclusion of the seven larvae that pupated after 96 h that gives an indication that the 2.8, n.s.). Without cleaning the arena between larvae there was a ADH to adult eclosion is reduced by this interruption in the 2nd distinct attraction to the quadrant to which the first larva instar stage. The impact of this disruption is more clearly travelled (Chi-square = 17.1, p < 0.001). indicated in Fig. 5, which shows a significant decrease in adult Where a previously generated trail was present, made by size with increased time away from the food source in the 2nd confining a number of larvae between two temporary walls of instar stage (F = 3.97, p < 0.003). Again the ANOVA is aluminium foil across the arena, then the experimental larvae strongly influenced by the small size of the 96 h cohort of placed in the centre of the arena showed a significant attraction adults, with Fisher’s PLSD post hoc test indicating no to the pre-laid trail compared with being ‘off’ trail (Chi- significant size difference between pair-wise comparisons of square = 11.1, p < 0.001). the adults from other groups. The simple behavioural observations (experiments 3 and 4) 4. Discussion were designed to see whether an individual larva’s dispersal is influenced by other larvae, by reacting to any chemical trails The estimation of the post mortem interval is frequently that might be present. Such behaviours on the one hand might calculated using ADH or ADD methods. In passing it might be reduce the time spent in locating suitable pupariation areas but observed that the use of the former is to be preferred over the conversely would lead to very high densities of pupae in latter because of its greater sensitivity to diurnal temperature particular areas with possible predation/parasitisation con- fluctuations. ADH calculations are invariably calculated for a sequences. specific species from laboratory studies and these constrained Where the larvae were individually started in the centre of and simplified conditions often do not reflect features that are the arena and the surface was cleaned between each larval run common in the real world. Ames and Turner [6] showed the the larvae took off in random directions and there was no non-intuitive effect of short cold episodes (which would also significant attraction to one quadrant or another (Chi-square = include the cold storage of larvae recovered from a crime scene prior to being sent or examined by an entomologist) on ADH calculations and this present study considers another area where there is a potentially frequent difference in the laboratory and field conditions, namely the ready availability of a suitable pupariation site very close to the corpse. Calculations of adult emergence ADH, based on laboratory studies, do not normally consider the post-feeding larval dispersal time separately. It is usually included in the total larval ADH (as for example in Table 1 in [6]), and in any case the dispersal time is usually short as the larvae leave the food and burrow into the adjacently provided pupariation medium. This present study suggests several areas that need to be considered in ADH calculations. The length of the post-feeding dispersal stage is important not only for the time it takes, but also for its impact on the following pupal development stage. Fig. 4. Impact on the total ADH to eclosion in Calliphora vicina of the removal No single value can be applied as the dispersal time will depend of 2nd instar larvae from their food for differing periods of time. The error bars on the dispersal behaviour of the specific blowfly species and are Æ1.96 S.D. from the mean. the terrain where the corpse was positioned. This study has 166 S. Arnott, B. Turner / Forensic Science International 177 (2008) 162–167 highlighted an area of the blowfly lifecycle that can introduce Acknowledgements errors into the use of insect development beyond the larval stage to estimate PMI. Further studies are needed for each of the main We thank Carole Ames for her input to this study. We would species to relate post-feeding dispersal to both temperature and also like to thank the two anonymous reviewers of the initial terrain so as to provide a more accurate estimate of the actual manuscript for their valued comments. This study was carried ADH and therefore PMI. out by S.A. in partial fulfilment of her M.Sc. in Forensic The experimental arena was of smooth plastic and whilst this Science at King’s College London. might be similar to vinyl or polished wooden floors it is quite different to the rougher textures of carpeting or out of doors References surfaces. In these latter cases it might be expected that greater energy expenditure would be needed by larvae to cover a [1] P. Nuorteva, Sarcosaprophagous insects as forensic indicators, in: C.G. particular distance and that progress generally would be Tedeschi (Ed.), Forensic Medicine a Study in Trauma and Environmental slowed. Hazards, vol. 3, Saunder, Philadelphia, 1977, pp. 1072–1095. Pupal duration, in individuals that had a long post-feeding [2] R.C. Bassanezi, M.B.F. Leite, W.A.C. Godoy, C.J. Von Zuben, F.J. Von Zuben, S.F. Dosreis, Diffusion model applied to post-feeding larval dispersal is lengthened relative to those with a short post- dispersal in blowflies (Diptera Calliphoridae), Memorias Do Instituto feeding dispersal period; and the resultant adults are reduced in Oswaldo Cruz 92 (1997) 281–286. size. This becomes a significant feature if the post-feeding [3] A. Gomes, C.J. Von Zuben, Postfeeding radial dispersal in larvae of larval dispersal is longer than 5 h and is likely to be the result of Chrysomya albiceps (Diptera; Calliphoridae): implications for forensic the loss of resources during the dispersal period. This finding entomology, Forensic Sci. Int. 155 (2005) 61–64. [4] J.W. Tessmer, C.L. Meek, Dispersal and distribution of Calliphoridae might not be of any relevance to scenes of crime in moist (Diptera) immatures from animal carcasses in Southern Louisiana, J. Med. woodlands but if the soils are hard and dried out, or Entomol. 33 (1996) 665–669. alternatively the scene of crime is indoors, then the larvae [5] C.J. von Zuben, R.C. Bassanezi, S.F. dosReis, W.A.C. Godoy, F.J. von may well spend considerable periods of time, counted in days, Zuben, Theoretical approaches to forensic entomology.1. Mathematical finding somewhere to pupate. Robinson [11] gives figures of 5– model of postfeeding larval dispersal, J. Appl. Entomol.-Zeitschrift Fur Angewandte Entomologie 120 (1996) 379–382. 14 days for the post-feeding dispersal period for C. vicina [6] C. Ames, B. Turner, Low temperature episodes in the development of which would add very considerable error to the PMI blowflies: implications for postmortem interval estimation, Med. Vet. calculations. Long post-feeding dispersal uses resources that Entomol. 17 (2003) 178–186. would otherwise have produced normal sized adults. Wild [7] L.G. Higley, N.H. Haskell, Insect development and forensic entomology, caught C. vicina do show a considerable range in adult size, in: J.H. Byrd, J.L. Castner (Eds.), Forensic Entomology: the utility of arthropods in legal investigations, CRC Press, Boca Raton, 2001. compared with the much more constant C. vomitoria. Further [8] A.A. Green, The control of blowflies infesting slaughterhouses I. Field experimentation is needed before adult size in C. vicina can be observations on the habits of blowflies, Ann. Appl. Biol. 38 (1951) 475–494. related to dispersal times or competition or partitioned between [9] C.-C. Hwang, B. Turner, Spatial and temporal variability of necrophagous them. Diptera from urban to rural areas, Med. Vet. Entomol. 19 (2005) 379–391. Previous studies have suggested that larvae dispersing from [10] S. Ireland, B. Turner, The effects of crowding and food type on the size and development of Calliphora vomitoria, Forensic Sci. Int. 159 (2006) 175– a food source is random [3] and the simple experiments 181. described here support this for individual C. vicina larvae. [11] W.H. Robinson, Urban Insects and Arachnids—A Handbook of Urban However, this does not hold for sequentially leaving larvae; Entomology, CUP, Cambridge, 2005. later ones are strongly influenced by the trails of those that have [12] M.P. Hassell, The Dynamics of Predator–Prey Systems, Princeton Uni- gone before them. The evolutionary value in following a trail is versity Press, New Jersey, 1977. [13] R.J Taylor, Predation, Chapman & Hall, London, 1984. not clear since the trail of one larva may not lead to a good [14] Y.Z. Erzinclioglu, Blowflies, Richmond Publishing Co. Ltd., Slough, pupariation site. Such behaviour is seen in larvae dispersing in 1996. huge numbers from human corpses at the Anthropological [15] C.N. Coyler, C.O. Hammond, Flies of the British Isles, Fredericke Warne facility, Tennessee (Amoret Whitaker, personal communica- & Co. Ltd., London, 1968. tion) and leads to areas of soil with very high densities of [16] E.N. Richards, M.L. Goff, Arthropod succession on exposed carrion in three contrasting tropical habitats on Hawaii island, Hawaii, J. Med. puparia. Such high densities may increase the chance of Entomol. 34 (1997) 328–339. survival although it is well known that high prey density attracts [17] W.C. Rodriguez, W.M. Bass, Insect activity and its relationship to decay high numbers of predators and parasitoids [12,13]. rates of human cadavers in East Tennessee, J. Forensic Sci. 28 (1983) 423– Differences between the observed and expected ADH values 432. (from Fig. 2) indicate that considerable errors may be [18] T.I. Tantawi, E.M. Elkady, B. Greenberg, H.A. ElGhaffar, Arthropod succession on exposed rabbit carrion in Alexandria, Egypt, J. Med. introduced by a variable post-feeding dispersal period when Entomol. 33 (1996) 566–580. using this method to estimate PMI. Assuming the constant [19] B.V. Travis, E.F. Knipling, A.L. Brody, Lateral migration and depth of 20 8C conditions of the growth of these experimental pupation of the larvae of the primary screw-worm Cochliomyia amer- calliphorids, a migration lasting 6 h will give a total ADH icana, J. Econ. Entomol. 33 (1940) 847–857. that is would over-state the PMI estimate by 1.25 days, an 8 h [20] K.G.V. Smith, An introduction to the immature stages of British Flies, Handbooks for the Identification of British Insects, vol. 10, Royal dispersal time overstates by 1.5 days, and a 24 h dispersal by 2.1 Entomological Society of London, 1989, pp. 1–280. days. These experiments were not run for the extended period [21] J.B. Cragg, The natural history of sheep blowflies in Britain, Ann. Appl. of 5–14 days quoted by Robinson [11] for C. vicina dispersal. Biol. 42 (1955) 197–207. S. Arnott, B. Turner / Forensic Science International 177 (2008) 162–167 167

[22] K.G.V. Smith, Insects and Other Arthropods of Medical Importance, [24] B. Greenberg, Flies as forensic indicators, J. Med. Entomol. 28 (1991) British Museum of Natural History, London, 1973. 565–577. [23] B. Greenberg, Behavior of postfeeding larvae of some Calliphoridae [25] K. Tullis, M.L. Goff, Arthropod succession in exposed carrion in a and a muscid (Diptera), Ann. Entomol. Soc. Am. 83 (1990) 1210– tropical rainforest on O’ahu Island, Hawaii, J. Med. Entomol. 24 1214. (1987) 332–339.