Looking into the puparium: Micro‐CT visualization of the internal morphological changes during metamorphosis of the blow fly, Calliphora vicina, with the first quantitative analysis of organ development in cyclorrhaphous dipterans

Abstract Metamorphosis of cyclorrhaphous flies takes place inside a barrel‐like puparium, formed by the shrinking, hardening and darkening of the third‐instar larval cuticle. The opacity of this structure hampers the visualization of the morphological changes occurring inside and therefore a full understanding of the metamorphosis process. Here, we use micro‐computed tomography (micro‐CT) to describe the internal morphological changes that occur during metamorphosis of the blow fly, Calliphora vicina Robineau‐Desvoidy 1830 (Diptera: Calliphoridae) at a greater temporal resolution than anything hitherto published. The morphological changes were documented at 10% intervals of the total intra‐puparial period, and down to 2.5% intervals during the first 20% interval, when the most dramatic morphological changes occur. Moreover, the development of an internal gas bubble, which plays an essential role during early metamorphosis, was further investigated with X‐ray images and micro‐CT virtual sections. The origin of this gas bubble has been largely unknown, but micro‐CT virtual sections show that it is connected to one of the main tracheal trunks. Micro‐CT virtual sections also provided enough resolution for determining the completion of the larval‐pupal and pupal‐adult apolyses, thus enabling an accurate timing of the different intra‐puparial life stages. The prepupal, pupal, and pharate adult stages last for 7.5%, 22.5%, and 70% of the total intra‐puparial development, respectively. Furthermore, we provide for the first time quantitative data on the development of two organ systems of the blow fly: the alimentary canal and the indirect flight muscles. There is a significant and negative correlation between the volume of the indirect flight muscles and the pre‐helicoidal region of the midgut during metamorphosis. The latter occupies a large portion of the thorax during the pupal stage but narrows progressively as the indirect flight muscles increase in volume during the development of the pharate adult.

insects (Erezyilmaz, 2006). This classification can be nevertheless problematic (Haug, Haug, & Garwood, 2016) but, in general, the immature stages of most holometabolous (5endopterygote) insects radically differ from the adult stage in morphology, behavior, and ecology. The last larval instar of holometabolous insects transforms into an adult through the pupal stage, which undergoes substantial morphological changes through a more or less extensive histolysis of the larval tissues and subsequent histogenesis of the imaginal ones. Swammerdam (1669) distinguished, however, a fourth group differing from other holometabolous insects in that the insect "does not shed the [larval] skin, but acquires the form of a Nymph under it" (pp. 17-18). This group roughly corresponds to the cyclorrhaphous dipterans, where the pupal stage and the following development of the adult certainly take place inside a barrellike puparium, formed from the hardening and darkening of the thirdinstar larval cuticle (Fraenkel & Bhaskaran, 1973;Martín-Vega, Hall, & Simonsen, 2016). This feature, found in only a few other insects (e.g., Strepsiptera, some Hemiptera), allows an extensive and complete histolysis of most larval tissues, as the insect lies within the rigid, protective puparium. Cyclorrhaphous flies include the insect model organism par excellence, the fruit fly Drosophila melanogaster Meigen, so the metamorphosis of this species has received special attention. However, recent studies of morphology have been few and the studies of Robertson (1936) and Bainbridge and Bownes (1981) still stand as two of the most detailed morphological analyses on the metamorphosis of D.
Although far less studied than D. melanogaster, the blow fly Calliphora vicina Robineau-Desvoidy (5 C. erythrocephala (Meigen)) has also been used as a model organism for the morphological study of insect metamorphosis, especially in the late 19th and the 20th centuries (e.g., Bautz, 1971;Lowne, 1892;P erez, 1910;Pihan, 1968;Possompès, 1953;Wolfe, 1954). Calliphora vicina is a widely distributed, synanthropic species of high economical (Aak, Birkemoe, & Leinaas, 2011) and medico-legal (Donovan, Hall, Turner, & Moncrieff, 2006) importance. Indeed, at the present time, studies into the metamorphosis of C. vicina and other blow flies have seen a renewed interest within a forensic context, as they are typically associated with decomposing organic matter. Staging the intra-puparial period allows the determination of age-specific morphological landmarks which, when applied to blow fly puparia collected from a forensic scene, can aid minimum post-mortem interval estimations (Brown, Thorne, & Harvey, 2015;Richards et al., 2012;Zajac & Amendt, 2012). However, despite its biological significance, and its applied importance in the particular case of blow flies, the internal morphological changes taking place during the metamorphosis of cyclorrhaphous flies are still poorly understood.
Moreover, as a consequence of the lack of models of morphological changes in cyclorrhaphous fly metamorphosis, there is a frequent confusion of the concepts in the entomological literature, which might lead to significant errors in applied studies .
In recent years, the use of X-ray micro-computed tomography (micro-CT) and computer-based 3D-reconstructions in zoological research has revitalized and revolutionized morphological and developmental studies, enabling the acquisition of high quality data of complex internal structures (Lauridsen et al., 2011;Smith et al., 2016). Furthermore, in contrast to histological techniques, the use of micro-CT does not require an invasive and time-consuming dissection of the sample (Carbayo & Lenihan, 2016;Simonsen & Kitching, 2014;Smith et al., 2016). Within a forensic context, Richards et al. (2012) demonstrated the potential of micro-CT for qualitatively describing internal morphological changes during the metamorphosis of C. vicina at 25% time intervals of the total intra-puparial period. Their preliminary results strongly supported the possibility of establishing a more accurate temporal resolution in further studies. Lowe, Garwood, Simonsen, Bradley, and Withers (2013) demonstrated how micro-CT can be used to yield volume measurements of selected organs and systems for ontogenetic analyses in a study on the metamorphosis of the painted lady butterfly Vanessa cardui (L.). Quantitative data are of particular interest in cyclorrhaphous flies as it has been stated that the insect volume is constant inside the rigid puparium during metamorphosis, in spite of the extensive histolysis and histogenesis which are taking place (Possompès, 1953). From pupariation and until the eversion of the head in the phanerocephalic pupal stage, that is, during the period when most larval tissues degenerate, a compensation mechanism for maintaining a constant volume is the development of a gas bubble which progressively increases in size within the apoptotic larval tissues in the abdominal region (Langley & Ely, 1978), although the origin of this bubble remains unclear (Denlinger & Zd arek, 1994). Despite these recent advances, there is still a lack of quantitative data on the rate of development of the organ systems of the adult fly.
The present study builds on the previous work by Richards et al. (2012), using micro-CT to describe the morphological changes taking place during cyclorrhaphous fly metamorphosis at a greater temporal resolution than anything hitherto published. The qualitative analysis of the internal morphological changes by Richards et al. (2012) was performed at 25% time intervals of the total duration of the intra-puparial period (i.e., from pupariation to adult emergence). Our aim is to refine the available temporal resolution to 10% time intervals of the total duration of the intra-puparial period, and down to 2.5% time intervals during the first 20% interval of the intra-puparial period, that is, the interval of major morphological changes .
Moreover, we aim to provide for the first time quantitative data on the development of different organ systems during the intra-puparial period. We hope that this study will not only lead to a better understanding of the morphological changes behind an evolutionarily critical process but also facilitate comparative studies of metamorphosis among different holometabolous and between holometabolous and non-holometabolous insect groups.

| Insect culture and sampling
A laboratory colony of Calliphora vicina Robineau-Desvoidy 1830 was established from adults collected using a modified Redtop® fly trap (Miller Methods, Pretoria) in the Wildlife Garden of the Natural History Museum, London. Newly-emerged adults from the colony were maintained at a controlled room temperature (238C 6 28C) and a daylight cycle of 18:6 hr (light:dark), to prevent the experimental population from entering diapause as post-feeding larvae (Saunders, 1987;Richards et al., 2012). The flies were provided with sugar, milk powder and water ad libitum during one week, and then also with 2 ml of pig blood (from pig liver) once daily during the following ten days, as a protein source for egg maturation. The flies were subsequently starved for 4-5 days, to permit adequate time for egg development. Finally, fresh pig liver was provided as oviposition medium.
Once the flies oviposited, the liver with the eggs was transferred to a plastic box (160 3 160 3 86 mm) containing an approximately 3 cm layer of autoclaved soil, and placed into an incubator at a constant temperature (248C 6 0.88C) without light, following standard protocol for blow fly rearing (Donovan et al., 2006). The larvae hatching from the eggs were reared in the same incubator and provided with additional small pieces (c. 15 3 5 cm) of fresh pig liver as needed. The feeding stage is comprised by three larval instars, each separated from the previous stage by a cuticular moult (Donovan et al., 2006). Once the post-feeding larvae started to wander from the food, the box was checked every 6 hr and the white prepupae, that is, irreversibly contracted third-instar larvae (Fraenkel & Bhaskaran, 1973), were placed into separate plastic boxes (120 3 120 3 60 mm) containing an approximately 1.5 cm layer of autoclaved soil and labeled with the pupariation time. A recent experiment showed that C. vicina larvae preferred soil as the substrate for pupariation although there were no differences in the total duration of the intrapuparial period among different substrates (Hartmann, Martín-Vega, Hall, & Amendt, 2016 vicina metamorphosis using temperatures of 24-258C (Bautz, 1971;Pihan, 1968;Possompès, 1953;Richards et al., 2012;Wolfe, 1954; Zajac & Amendt, 2012).
The remaining five puparia of each batch of 10 were kept as a reserve, and some of them used for histological studies later (see below). Each puparium was pierced in three places using an insect pin (in head, thoracic, and abdominal segments) to enhance the penetration of the staining solution. They were stained by immersion in a 0.  1910;Robertson, 1936) and no quantitative data are available. On the other hand, Richards et al. (2012) suggested that the development of the indirect flight muscles, from just short fibres in the first quarter of the intra-puparial period to occupying almost the entire volume of the thorax at the end of the fourth quarter, might be highly ageinformative if a quantitative measure of this organ system could be achieved. Segmentation was performed automatically using the "Magic wand" tool after redefining the grey scale range within each particular region of interest. The segmented volumes were then reviewed slice by slice and completed with manual segmentation where needed.
Quantitative data were obtained using the "Material statistics" module.
As some sections from the foregut and the helicoidal region of the midgut were difficult to segment accurately due to lack of contrast, Volume measurements were transformed logarithmically on both axes and a simple regression was calculated using the least squares method to describe the relationship between cross-sectional volume data from the indirect flight muscles and the pre-helicoidal midgut through development. The regression line was therefore defined as the power law equation log(y) 5 b 1 k * log(x), where k is the allometric coeffient.

| Histological studies
Additional puparia were subjected to histological studies to corroborate the observations from the micro-CT virtual sections, with a special emphasis on the pupal-adult apolysis event (i.e., the separation of the epidermal cells of the adult from the pupal cuticle). Four of the 15 remaining puparia (i.e., 5 non-scanned from each replication) were collected randomly for histological studies at 48, 120, 168, and 240 hr after pupariation (i.e., 20%, 50%, 70%, and 100% of the total intrapuparial period). The puparial case was removed from each specimen, and the insect was cleared with butanol, embedded in paraffin and sectioned in 10 lm thick sections on a Leica Reichert-Jung 2040 microtome. The resulting sections were stained with Weigert's haematoxylin, bluish erythrosine, phosphomolybdic acid and fast green and mounted on microscope slides in DPX. Photographs were taken using a Leica® DM6000 B microscope.

| 2D X-ray study
The development of the internal gas bubble during early metamorphosis was further investigated with 2D X-ray imaging using the X-ray beam of the micro-CT scanner. A new batch of C. vicina eggs were reared under the same conditions as described above (constant temperature of 248C 6 0.88C, the white prepupa considered time zero). At 0, 3,4,6,13,18,24,25,26,27,28,29, and 30 hr after pupariation, 9-10 puparia were collected and stuck by double sided adhesive tape to a Petri dish, approximately 2 mm apart and divided between two rows.
The Petri dish was then mounted horizontally on a polystyrene foam  3.1 | Prepupa 3.1.1 | 0% of the total intra-puparial period Unlike most other larval organs, the brain ( Figure 1c) will persist into the adult stage (Hartenstein, 1993). Regrettably, in this and subsequent development intervals, the edges of some neuropils-where neuronal processes contact and form synaptic connections (Ito et al., 2014)were usually blurry and not well defined. Although iodine staining has been proved to be suitable for analyzing insect neuroanatomy with micro-CT scanning (Sombke, Lipke, Michalik, Uhl, & Harzsch, 2015), it can result sometimes in more blurred neuropil edges and poorer contrast thresholds in comparison to other staining solutions as phosphotungstic acid (PTA) (Smith et al., 2016). An ongoing study using PTA as the staining method (Smith et al., 2016) is focussing on the reorganization of the brain and eye development during metamorphosis, and therefore few details on these structures will be discussed here.

| 2.5% Of the total intra-puparial period
Six hours after pupariation (i.e., 2.5% of the total intra-puparial period), the larval-pupal apolysis (i.e., the separation of the epidermal cells of the pupa from the larval cuticle or puparium) is well in progress, albeit at different levels in different body regions: it is nearly complete in the thoracic region, but has only started at some sections of the abdominal region ( Figure 1d-h). Contemporaneously to the larval-pupal apolysis, the extensive histolysis of the larval tissues has started, and the respiratory horns evert and push against the puparial wall ( Figure 1f). Also, the scans show a small gas bubble occupying space within the apoptotic larval tissues in the abdominal region ( Figure 1d-h), where the larval midgut still has a tubular appearance.

| 5% Of the total intra-puparial period
Twelve hours after pupariation (i.e., 5% of the total intra-puparial period), the larval-pupal apolysis is complete in the thoracic region and for the most part of the abdominal region (Figure 1g,h). Furthermore, the gas bubble has increased in volume, whereas the larval midgut Larval foregut visible in the anterior body region.
Larval midgut visible as a long coiled tube.
Larval hindgut visible as a long tube.
Larval salivary glands visible in the anterior body region.
Not developed.
24 (10%) Figure 3c  The correspondent percentage of time of the total intra-puparial period (IPP) is given in brackets after each time. , medial cross section of the abdomen. br, brain; cps, cephalopharyngeal skeleton; gb, gas bubble; lfg, larval foregut; lhg, larval hindgut; lhm, larval hypodermal muscles; lmg, larval midgut; ltt, larval tracheal trunks; lsg, larval salivary glands; rsph, respiratory horns. Arrows indicate sites where larval-pupal apolysis has occurred appears to have begun to contract (Figure 1g,h). Indeed, X-ray observa- tions of the gas bubble shows that the bubble is already formed 3-4 hr after pupariation and then progressively increases in volume during the following hours ( Figure 2; Table 2 3.2 | Pupal stage 3.2.1 | 7.5% of the total intra-puparial period Eighteen hours after pupariation (i.e., 7.5% of the total intra-puparial period), the larval-pupal apolysis is complete as the epidermis is fully detached from the puparium (Figure 3a,b), although the main tracheal trunks are still attached to the posterior spiracles of the puparium (Figure 3b). At this point, the insect is no longer a prepupa and should be termed a cryptocephalic pupa (Fraenkel & Bhaskaran, 1973;Martín-Vega et al., 2016). In the cryptocephalic (5"hidden head") pupa, the cephalopharyngeal skeleton has been partially extruded and the legs and wings have partially everted, but the head remains retracted (Figure 3a,b). The larval midgut continues contracting (Figure 3a).

| 10% Of the total intra-puparial period
Twenty-four hours after pupariation (i.e., 10% of the total intra-puparial period) (Figure 3c-g), the adult midgut has fully contracted, forming a dense mass of apoptotic larval midgut cells, that is, the yellow body, overgrown by the continuous epithelial layer of the adult midgut (Hakim et al., 2010;Takashima et al., 2011). The adult midgut is sackshaped, closed at both ends and displaced to the ventral side of the abdomen by the gas bubble (Figures 3c,d,g and 4a), which has continued expanding (Figure 2; Table 2), now occupying the central part of the abdomen (Figure 3g). The Malpighian tubules, which according to Bodenstein (1950) separate from the larval gut and will persist into the adult stage, can be distinguished lying between the bubble and the adult midgut. The apoptotic larval hindgut is still present in the caudal part of the abdomen (Figure 3d,e). The volume of the gas bubble continues to grow until it reaches maximum size; it then remains more or less constant during the following hours. However, 27-30 hr after pupariation the volume rapidly decreases until the bubble disappears as the gas is released into the space caused by apolysis between the pupa and the puparial cuticle ( Figure 2; Table 2). The gas from the bubble actually escapes between the pupa and the puparium, at first to the posterior part of the puparium (likely through the closest tracheal trunk) and then surrounding the pupa (Figure 2a). It thus creates the necessary space in the anterior part of the puparium for the head of the pupa to evert by muscular contractions (Hall, Simonsen, & Martín-Vega, 2017).

| 12.5% Of the total intra-puparial period
Thirty hours after pupariation (i.e., 12.5% of the total intra-puparial period), the head, legs and wings have been fully everted (Figures 2a   and 5a,b) and, therefore, the cryptocephalic pupa has been transformed into the phanerocephalic (5"visible head") pupa (Fraenkel & Bhaskaran, 1973;Martín-Vega et al., 2016). Thus the brain is now located in the head, which is hyaline in appearance until haemocytes and fat bodies migrate from the body and fill it out (Figure 5a-d). As a consequence of the evagination of the head, the respiratory horns (Figure 5e) move backwards and will be projected to the outside of the puparium (see Greenberg, 1991 for more details) through the bubble membrane (Sukontason et al., 2006). Moreover, once the gas bubble has disappeared, the sack-shaped adult midgut expands and occupies the majority of the thorax and the anterior part of the abdomen (Figures 5a-g   and 4b). The abdomen significantly shortens after head eversion and shows an extensive histolysis of remaining larval tissues, such as the abdominal musculature and the hindgut (Figure 5b). No significant changes were observed during the following 6 hr, that is, up until 36 hr after pupariation (i.e., 15% of the total intra-puparial period), apart from a more advanced histolysis of the larval hindgut (Figure 5c).

| 17.5-20% Of the total intra-puparial period
The morphology of the pupa is very similar at 42 and 48 hr after pupariation (i.e., 17.5% and 20% of the total intra-puparial period) (Figure 6a (Figures 8c, 4c, and 6a,c); this is in accordance with Takashima  Table 3). Furthermore, the histogenesis of the indirect flight muscles has started and small fibres of both dorso-ventral and dorsallongitudinal muscles are present (Figures 6b, 8d, and 9a-c; Table 3).
According to Fernandes, Bate, and Vijayraghavan (1991), myoblasts surround modified larval muscles and use them as templates for forming the dorsal longitudinal muscles (Figure 7b), whereas the dorsoventral muscles develop simultaneously without the aid of such templates.

| Pharate adult
3.3.1 | 30% Of the total intra-puparial period Seventy-two hours after pupariation (i.e., 30% of the total intrapuparial period), the pupal-adult apolysis is complete over the entire body (Figure 10a-h) and the insect is therefore no longer a pupa but an adult, termed pharate adult as it is still within the puparium (Fraenkel & Bhaskaran, 1973;Hinton, 1946;Martín-Vega et al., 2016). The midgut now occupies the anterior half of the abdomen and the hindgut is a continuous tube, connecting between the midgut and the rectal pouch (Figures 4d and 10b). The midgut is still voluminous (Figure 8b; Table 3), and it is now stretched anteriorly, being distinctly bottle shape (Figures 4d and 10a-b). The bottleneck or stretched region of the midgut is located in the same anterior section of the thorax where the thoracic ganglion is now positioned and where the crop and the cropduct are developing (Figures 4d and 10a,b). The oesophagus is also developing, although the pre-and post-ganglionic sections still appear to be poorly developed and they are therefore difficult to segment by Avizo software (Figures 4d and 10a). The reproductive organs are also distinguishable at this time (Figure 10f-h). In the eyes, the ommatidia start to be discernible below the cornea (Figure 10e

| 40% Of the total intra-puparial period
At 96 hr after pupariation (i.e., 40% of the total intra-puparial period), the whole thoracic portion of the midgut is stretched (Figure 8a) as the crop-duct grows and the crop is positioned in the posterior region of the thorax; the pre-helicoidal region of the midgut thus acquires the shape of a long-necked bottle (Figures 4e and 11a). The posterior end of the midgut grows and becomes helicoidal (Figures 4e and 11a-b). In the head, small fibres of the dilator muscle of the pharynx can be observed at this stage (Figure 11a).

| 50% of the total intra-puparial period
The crop reaches its final position in the anterior region of the abdomen 120 hr after pupariation (i.e., at 50% of the total intra-puparial period), where it expands lateroventrally acquiring a flattened shape (Figures 4f and 11c). With the crop out of the thorax, the pre-helicoidal region of the midgut becomes more cylindrical, decreasing the diameter of its abdominal section and increasing the diameter of the thoracic section (Figures 8a, 9d-f, and 11c). Both the helicoidal region of the midgut and the hindgut grow and become elongate (Figure 4f), developing their typical left-right loops (Lengyel & Iwaki, 2002), thereby roughly acquiring the final shape of the alimentary canal of the adult (see Graham-Smith, 1934). The adult salivary glands also appear to be fully developed at this time (Figure 11d).

| 60% of the total intra-puparial period
From 144 hr after pupariation (i.e., 60% of the total intra-puparial period) until the end of the intra-puparial period, the diameter and the volume of the pre-helicoidal region of the midgut will decrease indirect flight muscles are attached to the thoracic cuticle-see Wisser and Nachtigall (1984) for details on the muscle insertion points-but during metamorphosis the fibres are mostly surrounded by haemocytes, fat bodies and fatty droplets (Figures 12a-e, 13ac, and 14a,b)-see Crossley (1965) for more details. The indirect flight muscles will only attach to their corresponding insertion points in the thoracic cuticle at the end of the intra-puparial period (Figures 12g, 13d, and 14c).

| 70-80% Of the total intra-puparial period
The final intervals of the intra-puparial period are also marked by a 3.3.6 | 90-100% Of the total intra-puparial period At 216 and 240 hr after pupariation (i.e., 90% and 100% of the total intra-puparial period, respectively), the yellow body has been completely absorbed and is no longer observable in the midgut, while the meconium fills the rectum and a considerable portion of the rectal pouch (Figure 12e-i). The rectal pouch goes through another significant increase in volume compared to the previous development interval (Figures 8c and 9g-i; Table 3).
In the head, the dilator muscle of the pharynx and the tentorial muscles have progressively developed from the small fibres present 96 hr after pupariation (Figure 11a) until its complete formation after 216-240 hr (Figure 12a (Finell & Järvilehto, 1983). The ptilinal invagination reaches its final extension at the end of the intra-puparial period (Figures 11d,f and 12a-h), when the pharate adult is ready to emerge from the puparium.

| The Metamorphosis of Calliphora Vicina
In cyclorrhaphous flies, the first signs of the degeneration of larval tissues are reported to appear immediately after pupariation (Levy & Bautz, 1985). This is supported by Cepeda-Palacios and Scholl (2000) who showed how larval-pupal apolysis also begins within 3 hr of pupariation in the sheep bot fly Oestrus ovis L. (Oestridae). However, this is in contrast to studies of other species such as Sarcophaga bullata Parker (Sarcophagidae) (Fraenkel & Bhaskaran, 1973) and C. vicina in which the apolysis was reported to start several hours after pupariation (see Table 4 for a compilation of the published data on apolyses timings of C. vicina). Nevertheless, our results show that the larval-pupal apolysis must start shortly after pupariation in C. vicina as well, given that it is nearly complete in the thoracic region only 6 hr after pupariation ( Figure 1d); that is, significantly earlier than hitherto described (Table   4). The timings of the apolyses in C. vicina determined by different studies are not fully concordant (Table 4), perhaps because of differences in the frequency of observations, with the data from Bautz (1971) being most in agreement with our observations. Determining when each apolysis is complete is fundamental to correctly establishing the duration of the different intra-puparial stages in developmental studies (Fraenkel & Bhaskaran, 1973;Martín-Vega et al., 2016) and our study considerably refines the timing of the apolysis events in C. vicina (Table   4). Furthermore, Cepeda-Palacios and Scholl (2000)  Among the morphological changes taking place during the metamorphosis of cyclorrhaphous flies, the full eversion of head, legs, and wings, which marks the transformation of the cryptocephalic pupa (Figure 3) into the phanerocephalic pupa ( Figures 5 and 6), are indisputably the most striking, radical and extensive (Figure 2a). It has been suggested that the development of the gas bubble in the abdominal region may serve not only for maintaining a constant body volume within the puparium during a period of extensive histolysis (Langley & Ely, 1978) but also for aiding in the eversion of head, legs and wings by creating enough space between the pupa and the puparium, once the gas from the bubble has been released (Bainbridge & Bownes, 1981;Zd arek & Friedman, 1986). The role of the gas bubble in the eversion of those  Relative volumes are given as a percentage of the total puparial volume.
structures is described in another study (Hall et al., 2017), but the bubble must indeed be critical for this process, as D. melanogaster mutants which fail to release the gas, result in either crypto-or micro-cephalic phenotypes (Rewitz et al., 2010). How the gas bubble forms is, nevertheless, still unknown. Zd arek and Friedman (1986) suggested that in the flesh fly S. bullata, it is formed by inflation of the midgut 15-20 hr after pupariation. However, in C. vicina, the gas bubble appears 3-4 hr after pupariation (Figures 1 and 2) and its volume increases before the adult midgut inflates (Figure 3c-d). Langley and Ely (1978) (Gilby & Rumbo, 1980;Zajac & Amendt, 2012).
Interestingly, once the gas bubble has been released and the head everts, the lumen of the adult midgut expands as if to occupy the space left vacant by the bubble (Figures 5a,d and 8a-b). Langley and Ely (1978) suggest that the development of the gas bubble may act as a compenzation mechanism enabling the insect to decrease its mass without decreasing its volume within the puparium during the prepupal and the cryptocephalic pupal stages, and they found a significant correlation between the bubble volume and the puparial weight loss in G.

morsitans.
Considering that there is a highly significant weight loss between the cryptocephalic and the phanerocephalic pupal stages in C.
vicina (Zajac & Amendt, 2012), our results suggest that the adult midgut may replace the gas bubble, while maintaining the body volume constant within the puparium during the phanerocephalic pupal stage.
Moreover, the significant negative correlation between the changes in volume of the indirect flight muscles and the pre-helicoidal region of the midgut (Figure 13) suggest that the development of both organ systems might be somewhat connected, with the indirect flight muscles progressively replacing the midgut as the largest organ within the thorax toward the end of the intra-puparial period (Figure 8b,d). There is evidence from other studies to support the hypothesis that organ growth in insects might be at least partly regulated by feedback between growing organs to modulate final sizes, including negative regulation (Stern & Emlen, 1999). Our results might be the first indirect evidence supporting this hypothesis in the metamorphosis of cyclorrhaphous flies. Nevertheless, the genetic architecture of insect body size appears to be complex, involving a large section of the genome (Carreira, Mensch, & Fanara, 2009

| Micro-CT as a tool for intra-puparial development studies
The use of micro-CT has a series of practical advantages over traditional histological studies as already highlighted by several authors (Carbayo & Lenihan, 2016;Lowe et al., 2013;Smith et al., 2016;Richards et al., 2012). Among the most important advantages, micro-CT is a significantly less time-consuming process and a much less invasive analysis of the sample which, subsequently, can be virtually dissected in any plane. Traditional histology is particularly challenging in the case of cyclorrhaphous fly puparia because of the profusion of fat bodies and fatty droplets (Davies & Harvey, 2013) resulting from the histolysis of larval tissues (Crossley, 1965), as well as the hardness and impermeability of the puparium, which makes the infiltration of the internal tissues with the embedding material difficult. Davies and Harvey (2013) tested different fixation and preservation protocols but in every case they obtained frequently fragmented sections, thus losing significant parts of the sample tissues. These drawbacks of histological studies can severely limit the morphological analyses and preclude obtaining quantitative data. In fact, those limitations can be extended to any method involving a sectioning of the sample, such as immunohistochemistry, transmission electron microscopy or serial block-face scanning electron microscopy, which moreover is limited to very small size samples (see   We must emphasise, however, that micro-CT should not be considered a complete substitute for other imaging methods for morphological analyses, but should rather be a complementary technique. As suggested by Saltin et al. (2016), the combination of the appropriate method and the right organism can provide new insights into specific questions on specific developmental processes. With the current available resolution, micro-CT scans cannot document changes at the cellular level (Richards et al., 2012). For example, monitoring the histolysis and histogenesis of individual hypodermal muscles at the cellular level during metamorphosis, where certain larval muscles transform into adult ones, requires the higher resolution power that can be delivered by other methods, like traditional histology (Crossley, 1965;Zajac & Amendt, 2012) or transmission electron microscopy (Takashima et al., 2008). Conversely, the muscles, as well as other tissues, require staining by both techniques and therefore each specimen can be measured only once during its development, losing information on potential individual variation in growth (Cock, 1966). The use of iodine in the micro-CT staining used here can result in overstained structures like the apoptotic larval salivary glands in the current scans (Figure 3c,d,g), probably due to a reaction between the iodine and the products from phosphatase hydrolysis taking place in cellular cytoplasm at that point (Levy & Bautz, 1985). There are a variety of staining solutions which can also produce high differential tissue contrast for X-ray imaging of tissues (Metscher, 2009). Further research on the suitability of different staining methods for specific tissues and optimal staining times (Smith et al., 2016) may overcome these drawbacks in the near future.  (Table 1). Also, the use of micro-CT enabled delimiting the start and completion of the apolysis events with higher precision (Table 4) We must highlight that the use of a software tool for automatic segmentation-the Avizo's "Magic wand" tool used hererequired in every case reviewing the virtual slices and manual correction by adding or removing areas where needed. For example, the indirect flight muscles usually showed contrast levels similar to the adjacent fat bodies, resulting in the inclusion of the latter in the automatic segmentation. However, this approach undoubtedly saves a considerable amount of time in comparison to a complete segmentation by hand when large data sets need to be analyzed. Different image analysis methods for micro-CT studies of insect internal morphology have been recently explored (e.g., Lowe et al., 2013;Simonsen & Kitching, 2014;Smith et al., 2016). Therefore, as with the choice of the imaging technique, certain image analysis methods will be appropriate for certain questions.

A U T H O R' S CONTRIBUTIONS
DMV, TJS and MJRH conceived and designed the study. DMV conducted the experiments, and analyzed the data. DMV, TJS and MJRH The actual start and completion times could be earlier due to non-continuous sampling.
MART IN-VEGA ET AL.