Early colonisation of urban indoor carcasses by blow flies ( Diptera : Calliphoridae ) : an experimental study from central Spain

Due to their ubiquity and synanthropy, blow flies (Diptera: Calliphoridae) are generally the first colonisers of cadavers and, therefore, frequently used to estimate a minimum postmortem interval (minPMI). Whereas in outdoor situations blow flies are expected to locate and colonise exposed cadavers within hours or even minutes after death, it is usually assumed that the colonisation of a cadaver indoors might be delayed for an uncertain period of time. This uncertainty severely limits the informativity of minPMI estimates based on entomological evidence. Moreover, these limitations are emphasised by the lack of experimental data on


Introduction
Blow flies (Diptera: Calliphoridae) are considered ideal forensic indicators due to their ubiquity, synanthropy and ability to locate and colonise cadavers promptly after death [1,2].
The colonisation by blow flies starts when the first female lays its eggs on the cadaver. Thus, when a cadaver is found, the time of insect colonisation can be accurately estimated by determining the age of the oldest immature specimens found on or around the body, if developmental data on the pertinent species are available and the environmental temperature of the forensic scene is modelled [1,3]. The estimated time of insect colonisation is then used to infer a minimum post-mortem interval (minPMI) rather than the actual time of death [1], as the latter should also include the pre-colonisation interval [4]. The length of this precolonisation interval will thus depend on how promptly the insects located and colonised the cadaver. In outdoor situations, blow flies are expected to locate and colonise exposed cadavers within hours or even minutes after death [1][2][3][4][5][6], so minPMI estimates can be very close to the actual PMI. On the other hand, oviposition might be delayed indoors, as it has been observed that fewer blow fly females are able to locate and access a cadaver under this condition [2,6]. It is therefore of pivotal importance to know which blow fly species are able to colonise indoor cadavers and how promptly, particularly if we take into account that most of the forensic cases involving entomological evidence have been reported to occur indoors [6][7][8][9]. However, the vast majority of studies on insect succession on carrion are performed outdoors, whereas available data from indoor locations come mainly from case reports [7][8][9][10]. Experimental data from indoor situations are very scarce [2,6], and therefore strongly needed.
The current study aims (i) to determine which blow fly species are able to colonise fresh carcasses placed indoors in an urban environment in central Spain, analysing potential changes in the species composition during the year [11,12]; (ii) to test potential differences in the earliest oviposition times between different blow fly species and seasons; and (iii) to examine if some areas on the cadaver are preferred for oviposition by the earliest colonisers.

Study site and sampling procedure
The experiment was conducted in a two-floor abandoned building (UTM 30T 472091, 4484682; 595 m a.s.l.) located in the centre of Alcalá de Henares, in central Spain (Fig. 1a).
Alcalá de Henares is a medium sized city of more than 200,000 inhabitants (data from the Spanish National Institute of Statistics), located ~30 km from Madrid, the capital of Spain.
The city area is located in the mesomediterranean bioclimatic level, characterised by hot and dry summers and humid and cold winters [12,13]. The abandoned building used for the experiment is part of the "Cuarteles de Lepanto y el Príncipe", an abandoned military complex (~43,000 m 2 ) built between 1859 and 1864. The complex currently belongs to the University of Alcalá, which is restoring it for academic use. The building used for the experiment was available for the study from June 2013 to July 2014 before being partially demolished for rebuilding. Inside the building, three empty rooms (20 x 20 x 3 m each) of the same characteristics were selected for the experiment. The rooms were contiguous but fully isolated, located in a long corridor, each room with one door facing northwest and one window facing southeast (Fig. 1a). The windows faced a courtyard and consisted of two fixed glass panes and two sliding glass panes (Fig. 1a), and two interior wooden-framed shutters (Fig. 1b). The doors were kept closed during the experiment, whereas the windows were left with one of the sliding panes and the interior shutters opened, leaving an entry area of 120 cm x 50 cm (Fig. 1a established when no more post-feeding fly larvae were observed dispersing from the carcasses; this and other aspects of the insect succession will be analysed in an upcoming publication. The pigs were killed by sodium phentobarbital overdose in accordance with the EU Council Directive 86/609/EEC and immediately transferred in plastic bags to the study site building, where they were unwrapped and exposed at 10:00 a.m on the day of death. No further ethical approval was needed in accordance with current regulations on animal experimentation (European Directive 63/2010/EU and Spanish Royal Decree 53/2013).
For every trial, one pig carcass was placed in each one of the three different rooms.
The carcasses were always placed in the same position: the head facing west and the belly facing the window (Fig. 1b). Each carcass was placed on an individual mesh platform (100 x 80 cm) in order to enable daily weight measurements (data not included here) and inspection and sampling under the carcass. Each platform was placed on a carpet of the same dimensions, in order to retain part of the decomposing liquids and fluids and facilitate the work around the carcass. Finally, each carcass was surrounded by a wooden strip quadrat (240 x 240 x 14 cm) affixed to the floor with silicone, in order to enable the daily collection of dispersing post-feeding larvae (data not included here). Hourly room temperature data was collected using calibrated data loggers; one logger was used per room.
The carcasses were visited 6, 24, 30, 48, 54, 72, 78, 96, 102, 120, 126 and 144 hours after exposure, and then every 24 hours until the end of each experimental trial (i.e. at 10:00 and 16:00 during the first 6 days and then daily at 10:00 until the end of the experimental trial). During every visit within the first month after carcass exposure, each carcass was visually and carefully inspected by two researchers in the search for egg batches. An egg batch was defined as a cluster of eggs oviposited either by one single female or by several females in aggregation [2]. Newly observed egg batches were recorded by photographs and on paper templates with schematic outlines of the pig bodies. Moreover, new egg batches were marked with nontoxic coloured powder (Fig. 1c), in order to not to record the same egg batch as new more than once. The place of each batch on the carcass was also recorded, distinguishing in this study the following seven oviposition sites: mouth, nostrils, ears, head

Data analysis
A principal component analysis (PCA) was performed to explore the relationship between the observed blow fly species composition and the different experimental trials. Moreover, similarity percentage (SIMPER) analyses based on the Bray-Curtis measure [14] were performed to identify which blow fly species primarily typified the carrion colonisation within each experimental trial. As some data did not met the assumption of normality according to Shapiro-Wilk tests, non-parametric Kruskal-Wallis analyses with post hoc Bonferroni were performed to test differences in the earliest recorded oviposition times (i) between the observed species within each experimental trial, and (ii) within the same species between different experimental trials, for those species occurring during more than one season. Finally, chi-squared tests followed by pairwise comparisons were performed to assess differences in the frequency of use of the different oviposition sites. For all analyses, differences were considered to be significant at the < 0.05 level. SIMPER analyses were performed using PRIMER v. 5 [14]; the other analyses were performed using Statgraphics Centurion (Statistical Graphics Corp. 1994-2000).

Results
The average temperature ± SD in each of the three rooms within each experimental trial were: activity and carcass decomposition will be analysed in an upcoming publication.

Species composition
In total, five blow fly species were observed to colonise indoor carcasses during the current study (Table 1) (Table 1). It must be mentioned, however, that C. vicina colonised only two of the three carcass replicates during summer-1 trial and was absent during summer-2 trial (Table 1). On the other hand, C. vomitoria and P. terraenovae only colonised the carcasses during the spring trial; C. vomitoria oviposited on the three carcass replicates whereas P. terraenovae oviposited only on two of them (Table 1). Finally, both Ch. albiceps and L. sericata colonised the three carcass replicates of summer, autumn and spring trials, but not the winter ones ( Table 1).
The PCA and SIMPER analysis supported the observed seasonal changes in the composition of the blow fly species colonising carrion during the current study, with the first two principal components accounting for more than 80 % of the variability (Fig. 2). In the PCA plot (Fig. 2), the points represent the scores of the observations (the experimental trials) and the vectors the weight of the variables (the recorded blow fly species) on the principal components. Points that are close together correspond to experimental trials with similar species composition, whereas vectors pointing in the same direction correspond to species with similar distribution among trials. Within each trial, the species contribution to the colonisation of the different replicate carcasses was very similar (Fig. 2), with average similarities of ~90 % or higher in all cases ( Table 2). Between trials, both summer experiments showed similar scores on the principal components (Fig. 2) and very similar species contributions, whereas the highest dissimilarity values were found between winter and both summer trials (Table 3). Both Ch. albiceps and L. sericata contributed approximately in the same proportion to characterise the carrion colonisation during summer, whereas the colonisation during winter was obviously typified by C. vicina alone as it was the only species occurring during that season ( Table 2). Calliphora vicina, Ch. albiceps and L.
sericata were the three main species characterising carrion colonisation during autumn and spring, although contributing at different levels in each season: in autumn, the three species contributed approximately in the same proportion to the colonisation of carcasses, whereas in spring C. vicina was clearly the species dominating the carrion colonisation ( Table 2). As mentioned, two additional species colonised the carcasses during spring: C. vomitoria and P.
terraenovae (Table 1). However, their contribution to the carrion colonisation during this season was very low, even negligible in the case of P. terraenovae (Table 2).
Overall, the number of observed egg batches was relatively low for every species and experimental trial with the exception of C. vicina in both winter and spring trials, when more than 20 egg batches were recorded on every replicate carcass at the end of each experimental trial (Table 1). Even if it is not possible to know how many females oviposited on each carcass, as some egg batches may correspond to several females ovipositing in aggregation, the low recorded numbers (Table 1) suggest that few females were able to locate and colonise the carrion. It must be highlighted, moreover, that the higher numbers of observed egg batches in winter and spring were due to the fact that carrion colonisation continued during most part of the first month after carcass exposure (Fig. 3) and even beyond: egg batches from C. vicina were observed until day 75 after carcass exposure on winter carcasses but not included in the analysis. In contrast, no new egg batches were observed during summer and autumn experiments beyond day 9 after carcass exposure (Fig. 3).

Earliest ovipostion times
Regarding the earliest oviposition times, C. vicina was the first coloniser in autumn, spring and on one carcass in summer-1, and the only coloniser in winter, laying eggs within the first 24 hours in most cases. The exceptions were the spring carcasses, where the first egg batches were observed 30 hours after carcass exposure in the three replicates, and one replicate carcass in winter, where the first egg batch was observed 168 hours after carcass exposure (Fig. 4). Nevertheless, no significant differences were found in the earliest oviposition times of C. vicina between trials (H = 4.37; p > 0.05). C. vicina was also the first coloniser in all the carcasses where it occurred with one exception: one replicate carcass in summer-1 trial, where it was observed 48 hours after carcass exposure, after the colonisation of L. sericata within the first 6 hours after exposure (Fig. 4). L. sericata was, besides C. vicina, the only other species able to colonise carcasses within the first 24 hours, although this was only observed in two of the replicate carcasses of summer-1 trial (Fig. 4). In general, L. sericata colonised the carcasses within 48-72 hours after carcass exposure in summer and autumn, whereas its arrival was delayed until 126-192 hours after carcass exposure in spring (Fig. 4).
There were, however, no significant differences in the earliest oviposition times of L. sericata between trials (H = 7.09; p > 0.05).
On the other hand, Ch. albiceps was clearly the last coloniser of carcasses in spring and summer, ovipositing 384-552 and 78-102 hours after carcass exposure, respectively, but it colonised the autumn carcasses significantly earlier (H = 9.51; p < 0.05), showing similar earliest oviposition times to L. sericata (Fig. 4). In accordance with all these observations, significant differences in the oviposition times by the different species were found within spring (H = 9.79; p < 0.05), summer-1 (H = 3.85; p < 0.05) and summer-2 (H = 3.97; p < 0.05) trials, with Ch. albiceps colonising carcasses significantly later and C. vicina and L. sericata being the primary colonisers in spring and summer, respectively (Fig. 4). In contrast, no significant differences between species were found within autumn trial (H = 5.77; spring (H = 9.79; p > 0.05). Finally, in the spring trial, C. vomitoria appeared to act as a secondary coloniser of carcasses, arriving 168-216 hours after carcass exposure, whereas P.
terraenovae showed an irregular colonisation pattern, with only two egg batches recorded 78 and 432 hours after carcass exposure, respectively (Fig. 4).

Oviposition sites
The mouth was the oviposition site where the first observed egg batch was most frequently found (χ 2 = 70.65; p < 0.05). Indeed, the mouth was preferred as the first site for colonisation in every experimental carcass with the exception of one of the replicate carcasses of winter trial, where the first egg batch was observed inside one of the nostrils. It must be emphasised once again that C. vicina was the first coloniser in virtually all the experimental carcasses where this species was present (i.e. all the experimental carcasses except one of summer-1 trial and the three of summer-2). Lucilia sericata was the first coloniser in those summer carcasses where C. vicina was absent, as well as in one of the summer-1 carcasses which C. vicina colonised later (Fig. 4). Taking into account the total number of egg batches at the end of the experiment, the three most frequent species (C. vicina, L. sericata and Ch. albiceps) used every body part as oviposition sites. However, C. vicina (χ 2 = 14.81; p < 0.05) and L. sericata (χ 2 = 15.15; p < 0.05) oviposited more frequently in natural orifices (mouth, nostrils, ears and anus), whereas Ch. albiceps laid its eggs more frequently on the trunk and legs (χ 2 = 61.6; p < 0.05).

Discussion
The five blow fly species recorded in the present study have been previously reported colonising human cadavers not only in Spain [10,15,16] but also across Europe and North America, including indoor situations [6][7][8][9][10]17]. Among them, C. vicina justifies its role as arguably the insect most widely used as a forensic indicator: it is commonly found across wild, rural and urban habitats [11,12,18], and the present results show that it might be able to colonise cadavers located indoors within the first 24-48 hours after death in every season ( Fig. 4), therefore potentially being the first species colonising a corpse under most situations.
Primary colonisation of indoor carcasses by C. vicina within the first 24-48 hours had also been reported by Reibe and Madea [2] using piglets during spring and summer in central Germany. In central Spain, C. vicina peaks in abundance during autumn and spring, whereas the number of active adults decrease drastically during summer at low elevations [11,12,18]. This explains why summer was the only season when this species did not colonise all the experimental carcasses (Fig. 4). On the other hand, C. vicina can be the only active blow fly species during winter [5,11,12,18], potentially monopolising the colonisation of carcasses during this season (Figs. 3 and 4). Domínguez Martínez and Gómez Fernández [15] reported how the colonisation of human cadavers by C. vicina alone was observed "over and over again" during winter in Spain, suggesting that the adults of this species may overwinter inside buildings and dwellings, thus enabling the rapid colonisation of carrion indoors even during the coldest periods of the year. A previous study conducted using carrion-baited traps in a nearby building [12] showed that C. vicina was also able to access carrion indoors during every season, even when doors and windows were closed, although it was absent in the hottest months of the year. Nonetheless, indoor carrion colonisation times under restricted access environments (i.e. no open doors or windows or other entry areas) still need to be investigated, as they may indicate a delayed access to the cadaver [9].
Although it shows a spatial and seasonal distribution similar to C. vicina [11,18], C.
vomitoria only colonised carcasses during spring in the current study, and significantly later than C. vicina in every carcass replicate (Fig. 4). Spring is, indeed, the season of maximum abundance for C. vomitoria in central Spain [11,18]. This species shows, besides, a less synanthropic character than C. vicina [19], which has been suggested as one of the reasons why C. vomitoria is less frequently observed colonising human cadavers [15]. Comparing the colonisation of indoor and outdoor piglet carcasses, Reibe and Madea [2] found C. vomitoria colonising only the carcasses placed outdoors. However, it must be taken into account that their study [2] was limited to the first 48 hours after carcass exposure, so later colonisations of indoor carcasses by C. vomitoria (Fig. 4) cannot be discarded. Velásquez et al. [20] found C. vomitoria colonising pig carcasses placed outdoors in a periurban environment in the south east of Spain during both autumn and winter, although in very low numbers and with a delay of 8-9 days after carcass exposure. Evidence therefore suggests that C. vomitoria is a secondary coloniser of carrion, under both indoor and outdoor conditions, so its presence on a human cadaver might be indicative of several days of exposure [20].
In their review of the insects collected in several indoor cases from Italy, Bugelli et al. [9] highlighted that Ch. albiceps and L. sericata were frequently found together colonising human cadavers. In our study, both species colonised every carcass replicate during both summers, autumn and spring trials (Fig. 4). Our results suggest, however, that Ch. albiceps tends to colonise indoor carcasses later than L. sericata; this was particularly clear during the spring trial (Fig. 4). Indeed, L. sericata and Ch. albiceps have been regarded as primary and secondary colonisers of carrion, respectively [5]. Nevertheless, the significant delay in carcass colonisation by Ch. albiceps recorded during the spring trial in the current study (Fig.   4) is more likely related to the scarcity of active adults of this species during early spring in the study area [12]. Although both species are only active during the warm months in the Iberian Peninsula, L. sericata adults show a longer period of annual activity, from early spring -"as soon as the warm weather can be felt" [15]to late autumn, whereas Ch. albiceps only peaks in abundance during the warmest months from mid-summer to early autumn [11,12,18]. It has been suggested that these different seasonal patterns might also represent an adaptive response by L. sericata to the well-known facultative predacious behaviour of Ch. albiceps larvae on the larvae of other blow fly species [11,20,21]. Although L. sericata may be able to colonise indoor carcasses within hours after death, it seems to locate them generally less promptly than C. vicina overall (Fig. 4). This agrees with the observations of Reibe and Madea [2] who observed L. sericata colonising indoor carcasses only after 48 hours of exposure, in a situation where C. vicina colonised them within the first 24 hours.
Regarding the occurrence of P. terraenovae, our results suggest that it might be just an exceptional coloniser of indoor cadavers in central Spain, given the low number of recorded egg batches (Table 1) and the irregular pattern of colonisation observed (Fig. 4). As a cold-climate species, P. terraenovae can be frequently found in central and northern Europe and North America, where it can certainly be one of the first colonisers of indoor cadavers [6,8]. However, the records of this species in the Mediterranean region are very scarce and it is usually absent in the carrion succession studies conducted in this area [22]. Interestingly, P.
terraenovae had only been recently reported colonising human cadavers in Spain: specifically, a single puparium was recovered from an indoor case in Madrid (central Spain) in spring, where the species was found together with C. vicina, L. sericata and Ch. albiceps immatures [22]. The occurrence of P. terraenovae in warm mesomediterranean locations might be related to the fact that its larvae are widely used as live bait for angling in Spain [22]; nonetheless, this needs further investigation.
As expected, the natural orifices of the cadaver, and particularly the mouth, were the first areas to be colonised by blow flies [1,20,23]. This is due to the fact that natural orifices offer both protection and soft tissues as mucous membranes for first-instar larvae to feed [23]. Natural orifices should therefore be carefully examined when searching for entomological evidence in fresh cadavers. As putrefaction advances and the epidermis detaches from the dermis [24], the oviposition on other body parts may be facilitated for later colonisers. The presence of gravid females and/or feeding larvae in a given area, moreover, can stimulate further oviposition [23,25,26]. During the feeding process, C. vicina and L. sericata larval masses move to the thoracic and abdominal regions of the body [20], so the observed preference of the secondary coloniser Ch. albiceps for those areas in the current study may respond to its aforementioned facultative predacious behaviour [20,21]. However, previous studies using carrion baits in Brazil did not found any clear preference of Ch.
albiceps for those baits already containing other blow fly larvae [27]. Further research on the interactions between Ch. albiceps and other coexisting blow flies, mainly L. sericata in Spain, is needed.
Finally, it is important to emphasise once again that it is not possible to know the exact number of egg-lays (i.e. oviposited by a single female) from our data. Blow fly females can oviposit in aggregations on carrion [26] and therefore each egg batch counted in this experiment (Table 1) might well have contained eggs from an undetermined number of females. Nevertheless, Reibe and Madea [2] reported significantly far higher numbers of egg batches on outdoor carcasses than on indoor carcasses for all tested exposure times. Anderson [6] also noted much fewer insects attending indoor carcasses than outdoor carcasses.
Cadaveric volatile organic compounds [28,29] released by outdoor cadavers are likely more easily detected, and perhaps females of some blow fly species are more efficient detecting them than others. Whatever the case may be, the low numbers of blow fly females attending indoor carcasses results in low larval densities feeding on carrion, and this results in turn in a slow decomposition process [2,6]. As noted by Anderson [6], carcasses placed indoors may thus provide suitable oviposition sites for much longer than carcasses placed outdoors. This was particularly clear during winter and spring trials in the current experiment (Fig. 3).
Nonetheless, a slow rate of decomposition can also be expected in carcasses placed outdoors during cool seasons [30,31].

Conclusion
Although delayed colonisation may occur, blow flies are able to colonise indoor carcasses There are obvious limitations to the placement of carcasses indoors that explains the paucity of studies on insect colonisation and succession on indoor carrion in urban environments [6]. Ideally, the present study would have included replicate buildings, an outdoor control pig and the completion of, at least, a second year of surveys, thus replicating every season. Regrettably, this was not possible due to the limited availability of the building.
Nevertheless, the current study should contribute to fill the existing gap in experimental data on insect colonisation of indoor carrion, particularly in the Mediterranean region.    Table 1.