Medical and Veterinary Entomology (2003) 17, 178–186
Low temperature episodes in development of blowflies: implications for postmortem interval estimation
C. AMES andB. TURNER Department of Life Sciences, King’s College London, U.K.
Abstract. Traditionallythe calculation of accumulated degree daysor hours (ADD or ADH) involves the concept of a minimum threshold temperature below which development ceases. Hence in fluctuating conditions, where tempera- tures drop below this threshold, there maybe periods of time when development is taken to be zero. This has important implications when the calculation of post- mortem interval (PMI) is based on the ADD or ADH of larval dipterans. Normal development of larvae of the blowflies Calliphora vicina Robineau-Desvoidyand C. vomitoria L. (Diptera: Calliphoridae) at 20 C was interrupted bycold episodes. The expectation was that total development time would increase bythe period at low (therefore no development) temperature but the total ADD or ADH should be the same as non-cold treated cohorts. The results, however, showed that total ADH for both species decreased linearlywith increasing temperature with no evidence of anyminimum threshold temperature effect. The increased ADH at low temperatures maybe due to either continued but reduced development or a delayin development restarting after the cold episode. Use of ADH in PMI estimations has shortcomings particularlyduring the winter period where low temperatures are involved or where there are sudden summer cold spells during the development period. As blowflydevelopment progresses from egg to pupa such errors will be compounded. Key words. Calliphora vicina, Calliphora vomitoria, development threshold tem- perature, forensic entomology, postmortem interval.
Introduction Northern Europe. In favourable conditions these species lay their eggs around natural orifices or wounds on the fresh In forensic science, insect evidence is most commonlyused corpse within a short time after death. These hatch to to estimate the time of death of a corpse (the postmortem larvae, which grow through three stages. During the interval, PMI). Some of the first insects to invade a corpse latter part of the third stage, larvae stop feeding, migrate in the U.K. are Calliphoridae (Diptera), including species of from the corpse and burrow down into the soil to pupate. Calliphora, Protophormia and Lucilia (Lane, 1975). In many The pupal stage has the largest duration before adults temperate regions Calliphora species are considered emerge. the most important, with Calliphora vicina (Robineau- Being poikilothermic, the rate of development in insects is Desvoidy) (¼ C. erythrocephala Meigen) and Calliphora governed byambient temperature; the higher the tempera- vomitoria (Linnaeus) being widelydistributed throughout ture the faster development occurs. Thus byknowing the ambient temperature and the progress of blowflydevelop- ment, an estimation of time since the eggs were laid can be made. Correspondence: Dr B. D. Turner, Department of Life Sciences, Two methods to estimate PMI using blowflydevelop- King’s College London, Franklin Wilkins Building, 150 Stamford St., ment are available. Larval size, usuallylength, can be London SE1 9NN, U.K. E-mail: [email protected] compared with data on larval size related to temperature
178 # 2003 The Royal Entomological Society Low temperature and blowfly development 179 and time in an isomegalendiagram (Grassberger & Reiter, room temperature (20–22 C). Flies were supplied with 2001). There are a number of difficulties with this approach, granulated sugar and water ad libitum. New flies were including the actual size, the limited number of published added to these populations as the experiment progressed. isomegalendiagrams and the restriction of this method to Each cage onlycontained approximately100 flies to avoid the larval stages only. The other approach is to measure the overcrowding (Saunders, 1997). accumulated degree days or hours (ADD or ADH) needed A few days before eggs were required, liquid liver exudate to reach a particular stage of development. Accumulated was provided as a protein source for the flies. Fresh pigs’ degree days or hours are the product of temperature above liver was then introduced as an oviposition and larval food a species’ (and possiblygeographical population) minimum source. Females massed upon the liver and oviposition developmental threshold and the time spent at that would normallybegin 30 min to 1 h later. The time of temperature. Prediction of insect development focuses on oviposition was noted and the liver and eggs were removed the range where the relationship between development rate from the adult cage. Eggs were then separated onto new and temperature is constant and is based on the work of liver in a disposable weighing boat. To prevent anylarval Abrami (1972), Allen (1976), Arnold (1960), and Baskerville & mass effects which might cause a localized temperature Emin (1969) (see also Wagner et al., 1984). One difficulty elevation (Vinogradova & Marchenko, 1984 cited in Catts & with this linear day/hour degree model is that it only truly Goff, 1992; Turner & Howard, 1992), eggs were separated applies to the thermal range where temperature is directly into small clumps of about two to five and spread over the proportional to development. The calculated ADD/ADH liver. The liver, which had been kept refrigerated, was required for development will be too low at temperatures allowed to warm up to room temperature before the eggs near the lower threshold and too high near the optimum were placed on it. Approximately30–40 eggs were used in temperature (Wagner et al., 1984). Additionally, although it each experimental replicate. is accepted that development ceases at low temperatures The weighing boats, containing liver and eggs, were (Laudien, 1973 cited in Lamb et al., 1984), the lower placed on a 2-cm layer of soil-less compost in individual threshold temperature is often verydifficult to determine plastic containers (7 Â 8 Â 11 cm) with a plastic lid. The accuratelybecause insects survive for long periods with near compost provided a medium in which the post-feeding zero development. The Development Threshold Tempera- larvae could pupate. The liver was changed regularlyso ture (DTT) is most commonlyestimated bymeasuring that food for the larvae was always in excess, to avoid developmental rates at a range of temperatures and fitting competition. The containers were then kept at 20 C (Cooled a regression line to the results. This can then be extrapolated Incubator, LMS, Sevenoaks, Kent, U.K.) and a second to the x-axis where development is zero. incubator (Cooled Incubator, Sanyo-Gallenkamp, U.K.) According to Higley& Haskell (2001) who reworked was used when needed for the alternative temperature. Kamal’s data (Kamal, 1958), the DTT for C. vicina and The laboratorywas air-conditioned at 20 C Æ 1 Cso C. vomitoria is 6 C, yet recent studies suggest this varies that the removal of larvae from the 20 C incubator for with locality(Saunders & Hayward,1998). The majorityof measurement did not varytheir temperature regime. The past research on the effects of temperature on development second incubator (monitored using a TinyTag, Gemini data on species in the Calliphoridae has been done at tempera- loggers, Chichester, U.K.) returned to the set temperature tures close to the optimum, mainlyover 20 C (Dallwitz, within 10 min of opening and closing the door to put in the 1984; Greenberg & Tantawi, 1993; Anderson, 2000). experimental larvae and was then left unopened until the Vinogradova & Marchenko (1984), however, did include end of the exposure period. Larvae were kept in continuous lower temperatures. In Britain and Northern Europe, dark. Statistical analyses were carried out using the SAS temperatures for the greater part of the year are relatively package STATVIEW v. 5.01. low. Calliphora vicina and C. vomitoria are able to develop at The effects of low temperature episodes were explored in these low temperatures and C. vicina is active through the year. two series of experiments on both blowflyspecies. Specific This studyreports on a series of laboratoryexperiments criteria for each series were influenced byprevious experi- that explore the effects of short periods of low temperature mental experience. Development was followed from the egg on the development of two locallycommon blowflyspecies stage through to adult eclosion. with particular reference to ADD/ADH estimation. A recent study(Myskowiak & Doums, 2002), considering the impact of the practice of refrigeration of insect evidential material Series 1 – Low temperature period at different development (using Protophormia terraenovae Robineau-Desvoidy) to stages temporarilyhalt development until passed to a forensic entomologist, is complementaryto this present study. Egg development took place at 20 C. Individual cohorts of larvae were then subjected to cold episodes, of 5 C for 5 days, at one of the following developmental stages – larval Materials and methods stages one, two or three (feeding) or the pupal stage. Following the cold episode the blowflies were returned to Wild populations of C. vicina and C. vomitoria were raised 20 C for the remainder of their development. Controls and maintained in gauze covered cages (22 Â 38 Â 27 cm) at remained at 20 C for the entire developmental duration.
# 2003 The Royal Entomological Society, Medical and Veterinary Entomology, 17, 178–186 180 C. Ames and B. Turner
Series 2 – Differing low temperatures for a constant time degree hours to move from stage to stage. Therefore, period regardless of where the cold episode occurred, whilst the developmental time will be longer (to make up the deficit of All experiments were kept at 20 C for 3 days following 1800 ADH), the ADH value to reach a specific stage should oviposition. Bythis time the larvae were late second stage. not be significantlydifferent from the corresponding con- Cohorts of these larvae were placed at 1 ,3 ,5 or 10 C for trol ADH value. 5 days, after which they were returned to 20 C for the This was not the case for the ‘post cold episode’ stages of remainder of development. Again, controls remained at either species (Fig. 2). Bartlett’s test for homogeneityof 20 C for the entire developmental duration. variances among treatments indicated that the group vari- Three replicate cohorts, each starting with approximately ances were not equal for either blowflyspecies ( C. vicina 30 eggs, were individuallyfollowed in each treatment within F ¼ 8.3, P < 0.001; C. vomitoria F ¼ 10.1, P < 0.001). As this an experimental series. Following oviposition, each cohort renders the standard one-way ANOVA inappropriate, Welch’s of eggs was checked half hourly, until all were hatched. one-way ANOVA test for data containing groups with Subsequentlytheywere observed every2 h between unequal variances was used instead. This showed that 10.00 hours and 18.00 hours each day. Different starting there are significant differences between the treatments for times were used to co-ordinate hatching and metamorphosis both species (C. vicina W ¼ 102.5, P < 0.001; C. vomitoria events with the daytime observational periods. Immature W ¼ 46.5, P < 0.001). This is seen in Fig. 2 bycomparing the stages were examined under dissecting microscope to deter- Control ADH values with the treatment ADH values where mine developmental stage. On pupariation, the puparia there is a difference of approximately600 ADH, suggesting were removed from within the compost and placed upon that no development took place in either species during the the surface. As the adults emerged, theywere counted and 5 C period: that is the DTT is around 5 C. removed from the containers. During all experiments, the The data were reanalysed using 5 C as the DTT. Welch’s time in hours since oviposition when the insects progressed test, applied to the recalculated ADH values in each treat- from one stage to another was recorded. ment, showed that for both species there were significant To calculate the ADH from the developmental time the differences between ADH values (C. vicina W ¼ 3.2, following formulae was used: P ¼ 0.016; C. vomitoria W ¼ 2.7, P ¼ 0.033). For C. vicina, the developmental stage at which the cold episode occurs ADH ¼ {[development time (h) À time in cold episode appears important. Scheffe´ ’s post hoc test (Table 2) shows a (h)]  20} þ [time in cold episode  temperature significant difference onlybetween the 1st stage larva and of cold episode ( C)]. pupal treatments, but low P-values between the 1st stage larval treatment and all the others for C. vicina. As a result of the uncertaintyof the concept of a DTT, This effect is not seen in C. vomitoria. Scheffe´ ’s test shows initiallyno correction for this was applied to anyof the data no significant differences between the treatments in this for either species. species. The ADH values to reach adult emergence are all similar to each other (P-values all > 0.1) regardless of when the 5-daycold episode occurs in the developmental period. Results
Development at 20 C – the controls Experimental series 2
Each of the experimental series included a set of controls As the temperature of the 5-daycold episode increased, at 20 C. The duration and cumulated values for the control the development time (in hours) to reach the adult stage for data have been combined in Table 1. At 20 C, the timings both species decreased linearly(Fig. 3). Both slopes are for the egg stage and 1st and 2nd stage larvae are quite significantlydifferent from zero (Fig. 3). There was an similar for C. vicina and C. vomitoria. The 3rd stage larvae increase of approximately8 h on the development time for period for C. vomitoria takes about 50% more time than for each C decrease in the 5-daycold episode. C. vicina, whilst the pupal stage is some 12% shorter If the DTT is 5 C (based on Experimental series 1), we (Fig. 1). Thus overall the total development time for would expect there to be no significant difference in the C. vomitoria is 34 h longer than C. vicina, 683 more ADH, developmental times for the 1, 3 and 5 C data. However, assuming both have a DTT of 0 C. a significant difference between the development times for each temperature was seen using Welch’s test on the data sets of each species (C. vicina W ¼ 2027.2, P < 0.001; Experimental series 1 C. vomitoria W ¼ 1529.4, P < 0.001). Using Scheffe´ ’s multiple comparison test on these data indicated significant differ- Excluding the controls, each treatment had a 5-dayperiod ences (P < 0.001) between all temperature combinations in at 20 C(¼ 2400 ADH) replaced bya 5-daycold episode of both species. This appears to be at variance with the find- 5 C(¼ 600 ADH). The linear hour-degree model suggests ings of the first experimental series (although this was that insects need to have accumulated a specific number of limited to onlyone temperature for the cold period) and
# 2003 The Royal Entomological Society, Medical and Veterinary Entomology, 17, 178–186 # 03TeRylEtmlgclSociety, Entomological Royal The 2003 eia n eeiayEntomology Veterinary and Medical
Table 1. Mean duration, standard deviation, range and accumulated degree hours of the egg, larval and pupal stages of Calliphora vicina and Calliphora vomitoria at 20 C.
C. vicina C. vomitoria
Mean Mean Mean Mean duration cumulative cumulative cumulative duration cumulative Cumulative Cumulative ANOVA of of stage duration duration duration of stage duration duration duration ADH C. vicina vs. Stage n (h) (h) SD min–max ADH n (h) (h) SD min–max ADH C. vomitoria Egg 160 24.7 24.7 1.3 21.8–22.8 493 113 23.6 23.6 1.2 20.5–26.3 472 ** Larval 1 97 17.9 42.5 4.9 36.3–50.5 851 150 15.5 39.2 4.2 35.0–47.5 783 **
, Larval 2 104 44.1 86.6 5.7 74.0–119.0 1733 103 47.2 86.4 5.8 70.0–96.5 1728 P ¼ 0.74 17
178–186 , Larval 3 140 123.4 210.0 12.9 179.0–252.8 4200 70 195.5 281.9 13.1 269.0–335.0 5638 ** Pupae 106 328.5 538.4 12.5 501.0–563.0 10769 111 290.7 572.6 18.1 540.0–636.0 11452 ** o eprtr n lwl development blowfly and temperature Low
**Significant at 1% significance level. 181 182 C. Ames and B. Turner
C. vicina a)
24.65 17.89 44.1 12000
ADH 11000 123.36 328.45 10000 Pupal Control Larval 1 Larval 2 Larval 3 Stage cold episode received
C. vomitoria b)
23.62 15.53 47.23 13000
195.5 12000 ADH 290.73 egg stage
lar va 1 11000 lar va 2
lar va 3 pupa Pupal Control Larval 1 Larval 2 Larval 3 Fig. 1. Comparison of relative proportions of the egg, larval and Stage cold episode received pupal stages to the total developmental time in Calliphora. vicina and Calliphora. vomitoria life cycles. Figures in the pie charts are Fig. 2. Boxplots of total ADH against the stage at which cold the mean duration of each developmental stage in hours at 20 C. episode (5 C/5 days) occurred. (a) Calliphora vicina,(b)C. vomitoria. *Indicates outliers. The line across the box is the median. The suggests there is no clear cut minimum development thresh- bottom of the box is at the first quartile (Q1), and the top is at the old for either blowflyspecies within the temperature range third quartile (Q3) value. The lines that extend from the top and covered in these experiments (1–20 C). bottom of the box show the range of the distribution. Rather than the expected consistent ADH values, across the range of temperatures (that is an expectation of gradient (C. vicina W ¼ 130.0, P < 0.001; C. vomitoria W ¼ 325.0, b ¼ 0), adult ADH values for both species follow a decreas- P < 0.001). Applying Scheffe´ ’s post hoc test to these data ing linear relationship with increasing temperature for the shows that the ADH values for the 1 C and 3 C treatments 5-daycold episode, which are significantlydifferent from are each significantlydifferent to those of all the other zero (Fig. 4). temperatures (Table 3). Using Welch’s test on each species’ ADH and tempera- Whilst statisticallythe ADH values for 5, 10 and 20 C are not ture data set (and taking the DTT point to be 5 C) reveals significantlydifferent (mean value for C. vicina ¼ 8057 ADH that, for both species, there is a significant difference and for C. vomitoria ¼ 8549 ADH), the ADH for the 1 and between 5-daycold temperature episodes and ADH values 3 Ctreatmentswerehigher(meanfor1 C cold period,
Table 2. Scheffe´ ’s multiple comparison significance table between total ADHs for differing treatments of cold episodes (5 C/5 days) during different life-cycle stages for Calliphora vicina (top right) and Calliphora vomitoria (bottom left). Bold indicates significant at the 5% level.
Larva 1 Larva 2 Larva 3 Pupa Control Larva1 0.050 0.134 0.009 0.088 Larva 2 0.973 0.960 >0.999 0.905 Larva 3 0.789 0.977 0.886 >0.999 Pupa 0.907 0.999 0.996 0.723 Control 0.943 0.417 0.105 0.175
# 2003 The Royal Entomological Society, Medical and Veterinary Entomology, 17, 178–186 Low temperature and blowfly development 183
a) a) 800 13000 12500 600 12000
y = –8.1251x + 699.17 2 11500 400 R = 0.8765 ADH ADH 11000
200 10500 y=–42.501x + 11583 R2 = 0.3269 10000 0 0 24681012 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 Temperature of 5-day cold episode (° C) Temperature of 5-day cold episode (°C) b) b) 13000 800 12500
600 12000
y = –8.1224x + 732.35 ADH 2 R = 0.9482 11500 400 y=–42.447x + 12247 ADH 2 11000 R = 0.5558 200 10500 0 24681012 14 16 18 20 0 Temperature of 5-day cold episode (°C) 0 2 4 6 8 101214161820 Temperature of 5-day cold episode (°C) Fig. 4. The effect of temperature of a 5-daycold episode on total ADH. Calliphora vicina: n ¼ 265, t ¼ 11.3, P < 0.001. (b) Calliphora Fig. 3. The effect of temperature of a 5-daycold episode on duration vomitoria: n ¼ 214, t ¼ 16.3, P < 0.001. of development to adult. (a) Calliphora vicina: n ¼ 365, t ¼ 43.2, P < 0.001. (b) Calliphora vomitoria: n ¼ 214, t ¼ 62.3, P < 0.001. DTT value affects the controls at 20 C because the hourly multiplier is the difference between the DTT and control temperature. Thus lowering the DTT from 5 to 0 C C. vicina ¼ 8619 ADH and C. vomitoria ¼ 9084 ADH; mean for increases the hourlymultiplier from 15 to 20. Lowering 3 Ccoldperiod¼ 8341 ADH and C. vomitoria ¼ 8823 ADH). the minimum threshold also changes the shape of the rela- As the DTT value was set at 5 C, the impact of the 5-day tionship and removes the obvious elbow point (interpretable period at 1 or 3 C should have had no numerical input to as some sort of threshold) seen when the DTT is set at 5 C. the ADH calculation. However, the data clearlyshows that the development rate outside the cold episodes was affected bythe low temperature experience. The cold episodes below the DTT had a negative influence, having some unknown Discussion effect, delaying development, resulting in an increased total ADH. Davies & Ratcliffe (1994) suggest that ADH values would The impact of the level set for the DTT on the ADH be expected to be approximatelysimilar for two congeneric estimate can be seen in Fig. 5. Two things are evident. As a species of similar size, such as C. vicina and C. vomitoria. result of the mathematics of calculating ADH, changing the The current studyfound that onlythe ADH to reach the 3rd
Table 3. Scheffe´ ’s multiple comparison significance table between total ADH for differing treatments of 5-daycold episodes at various temperatures during the late 2nd larval stage for Calliphora vicina (top right) and Calliphora vomitoria (bottom left). Bold indicates significant at the 5% level.
1 C3 C5 C10 C20 C 1 C <0.001 <0.001 <0.001 <0.001 3 C <0.001 <0.001 <0.001 0.004 5 C <0.001 <0.001 0.829 0.903 10 C <0.001 <0.001 0.996 0.440 20 C <0.001 <0.001 0.999 0.983
# 2003 The Royal Entomological Society, Medical and Veterinary Entomology, 17, 178–186 184 C. Ames and B. Turner