A variety of cell types develop when cells of h Drosophila embryos are cultured in an improved medium. Nerve, muscle, fat-body, chitin-secreting, and macrophage-like cells (possibly haemocytes) appear in the first 24 h and mature over the next week. Tracheal, imaginai disc, a second stage of the macrophage-like, and a number of unidentified fibroblastic and epithelial cells appear in the 2nd and 3rd week, following a resumption of cell multiplication. There is some organization of some of the cell types into higher structures.

The development of five cell types in in vitro cultures of early embryonic cells of Drosophila melanogaster was previously reported (Shields & Sang, 1970). Using an improved medium and a slightly modified technique we have now observed further developments in the cultures, involving more complete differentiation of some of these cell types and appearance of a range of other types. The five cell types were neuron-like, muscle, epithelial-type, macrophage-like, and fibroblast-type. Major further advances are shown by the epithelial-type and the macrophage-like cells, and the former are now believed to be fat-body cells and the latter seem likely to be haemocytes: the revised developmental cycle of both these types is given in detail. The new cell types which appear include chitin-secreting, tracheal, and imaginai disc cells, and a number of unidentified forms: only the first three of these are reported on in detail, the unidentified cells being simply noted as a group.

The main modifications of our earlier experimental procedure were omission of the trypsinization process, and the use of a new kind of culture slide. Thus, following dechorionation and surface-sterilization with hypochlorite solution, embryos were homogenized directly in the culture medium, washed twice with it to remove yolk and non-cellular debris, and the cells suspended in fresh medium before mounting. Dissociation, while not quite so complete as with trypsinization, was good, and the subsequent development of the cultures was much improved. The new slides consisted of two ordinary microscope slides stuck together with silicone rubber, one of them with a hole drilled through it to give a chamber 14 mm wide and 1 mm deep; the floor of this chamber was siliconized to give a hydrophobic surface. These slides greatly increased the stability of the culture drops, and made medium-changing an easier, and more precise, technique.

As before, cultures were mounted as column drops, which were prepared by placing a drop of cell suspension in the centre of the slide chamber and lowering a square coverslip onto it, sealed in place with vaseline. The drop was made sufficiently large to give a column about 5 mm in diameter, so that the air to medium volume ratio in the chamber was about 7·5:1. The slides were then held inverted for 1 h, to allow the cells to settle and attach to the coverslip, before incubating at 25 °C.

The greater depth of the new culture chambers meant that for a given settled cell density there was a greater volume of medium than in our earlier experiment, and to allow for this, medium-changing was performed about every 4 days rather than every 3 days. In practice, the frequency of medium changing depended on the rate of development of the cultures, which varied from experiment to experiment, and had to be judged from visual observation. As previously, changes were made by swivelling and lifting the coverslip to split the drop, removing old medium from the portion left in the chamber and replacing it with fresh medium. It was found advisable to restrict the first two changes to about a 25% replacement, rather than the 50% replacement used subsequently.

Cultures were initiated from embryos which were old at the time of dissociation, and Oregon S strain embryos were used.

The composition of the revised culture medium is shown in Table 1. The chief differences from the earlier medium were: (1) reduction of the Mg and Ca salts and of the yeastolate, (2) increase of the bicarbonate and glucose, (3) some alterations in the amino acid composition, (4) inclusion of choline chloride, and (5) omission of the organic acid salts. The bicarbonate correction should be emphasized, since other workers have said that this substance can be omitted from media for Drosophila embryonic cells. We have found that if yolk is thoroughly removed from cell preparations, and serum and yeastolate used at the levels shown, there is a definite requirement for bicarbonate. The amount employed in our previous medium was on the border-line of adequacy, and was therefore doubled.

Table 1

Composition of the new culture medium (mg/100 ml)

Composition of the new culture medium (mg/100 ml)
Composition of the new culture medium (mg/100 ml)

In making up the medium, the pH was again adjusted with 1% NaOH, sufficient being added to raise the pH to 6·6, and the KHCO3 was introduced at that point to bring it to the final value of 6·8–6·9. Loss of CO2 due to breakdown of the bicarbonate under acid conditions was thus minimized.

The added serum was usually non-inactivated, as before, since a slight lytic activity was then introduced which was helpful to the cultures in promoting separation, flattening-out, and multiplication of some of the cell types. Heat-inactivated serum was also tried, and gave essentially similar results. Completed medium was always held for at least a week before use, to reduce the lytic activity from a sometimes excessive initial level.

Despite the improvements in methods, the cells remained sensitive to culture in vitro, and showed great variability in the rate and pattern of their development. This variability probably derived from such factors as the degree of dissociation and damage caused during homogenization, the initial cell-density, the size and timing of medium changes, and the quality of the serum used in the medium. The following account outlines the pattern of development in the better cultures, with some mention of more important variations.

Nerve and muscle cells

The nerve (neuron-like) and muscle cells develop much as in the original preparations, and need no description beyond that already given (Shields & Sang, 1970). Both begin to appear between 6 and 12 h after mounting of the cultures – in some cases following cell division – and most have attained close to their final form by 24 h. Further development over the next week is of a relatively secondary character: it involves, in the case of the nerve cells, continued extension of their nerve fibres and a gathering of these into bundles which connect cell clusters throughout the cultures; and for the muscle cells, expansion of a proportion to sheet-like elements, and fusion of some of them to give simple syncitia. Detailed studies on both these cell types have been carried out by Seecof and co-workers (1971, 1973a, b), which establish their identity beyond reasonable doubt.

Fat-body cells

In our first paper the development of epithelial-type cells was described which were reported as mostly eventually rounding-up and filling with oildroplets. The latter change was thought to be a degenerative one, from its similarity to a process which occurs with other cells undergoing deterioration in culture, and the epithelial form was assumed to be the end-point of normal development. But under the improved culture conditions the change occurs more rapidly and proceeds further, and it is now seen as being a continuation of normal development towards the mature form of fat-body cells. Thus, the full sequence of development of these cells now appears as follows.

Most fat cells appear between 6 and 24 h of culture, as rather condensed, often fusiform elements, which derive by flattening down of earlier rounded cells. Flattening and expansion continue to between 24 and 48 h, to give the full epithelial form of large, flat, polygonal elements with cytoplasm richly-charged with dark granules and the nucleus prominently nucleolated (Fig. 1A). At 24–48 h most of them begin to round-up again, and rounding is completed by the third to fourth day. A variable number of cells do not flatten out, but remain rounded from the beginning, and some carry on flattening into the third day to give more circular forms with a wide border of clear cytoplasm which resemble the macrophage-like cells, but both these variants later continue developing in the normal way. The further extension of flattened forms into branched, granule-free elements described in our first paper now rarely occurs, and is believed to be an aberrant development.

Fig. 1

(A) Early fat-body cells, in flattened epithelial form, from 36 h culture. (B) Fully-mature fat-body cells from 10-day culture. Rounded, with large fatdeposits; cells integrating. (C) Early macrophage-like cells in flattened circular form from 3-day culture. (D) Multiplying early macrophage-like cells from 10-day culture. (E) Transformed, large macrophage-like cells from 18-day culture. All photographs of the same magnification in this, and subsequent Figures, unless otherwise indicated.

Fig. 1

(A) Early fat-body cells, in flattened epithelial form, from 36 h culture. (B) Fully-mature fat-body cells from 10-day culture. Rounded, with large fatdeposits; cells integrating. (C) Early macrophage-like cells in flattened circular form from 3-day culture. (D) Multiplying early macrophage-like cells from 10-day culture. (E) Transformed, large macrophage-like cells from 18-day culture. All photographs of the same magnification in this, and subsequent Figures, unless otherwise indicated.

The rounding cells go on enlarging, and at some stage numerous small, refractile oil- or fat-droplets begin to appear in their cytoplasm. This may occur as early as the second day, before rounding has actually begun, but more usually takes place on the third or fourth day, in fully-rounded cells. The droplets swell rapidly, further enlarging the cells, which by about the eighth day may be up to 4 times their initial rounded diameter, and packed with large globules. They are then much the most prominent cells in the cultures (Fig. 1B). However, the fat-filling phase is a particularly sensitive one, and there are frequently delays and blockages in both the cell-enlargement and fat-formation processes, leading to lateness and small size of the final product and abnormal forms. A range of variant derivatives is thus usually found, even in the best cultures. At about the eighth day, the majority of the fat-filled cells generally undergo a change, perhaps involving an alteration in their surface properties: they tend to become more loosely attached to the coverslip, to show changes of shape and position, and to integrate very closely with one another in groups of up to about 10 cells. At the same time many show a loss of their fat-contents (whether by re-absorption or by discharge into the medium is not known) and are left as dense, finely granulated elements with a very much wrinkled surface; and some vacuolate and disrupt. The latter effects are probably degenerative ones, perhaps resulting from an increased sensitivity of the cells to in vitro conditions, but the changes as a whole may reflect attainment of a further stage in the development of the cell type.

Those cells which do not lose their fat-deposits may show a slight further enlargement over the next few days, and become more translucent (seemingly due to a clearing of their cytoplasm) and are then virtually indistinguishable from mature larval fat-body cells. They show little further change after this, but tend to degenerate quickly, and to be lost by disruption and detachment at medium changes.

Macrophage-like cells

The new developments shown by the macrophage-like cells are: an early appearance of multiplying forms among them, and the later transformation of some of these multiplying forms to a more advanced product. As previously indicated, most cells appear initially as rather bulky, contracted, granular derivatives, tending to put out a frail membrane and/or fine pseudopodal processes. These elements arise within the first 24 h of culture, and flatten out irregularly over the next few days, losing some of their granulation as they do so, to give the typical thin, clear, circular forms described in our first paper (Fig. 1C). The circular forms can go on increasing in size, and still give the giant, multi-nucleate, and medusoid products originally reported; but the latter are now thought to be abnormal or degenerative developments.

The multiplying cells begin to appear among the normal circular-form cells at around the seventh day, as condensed, rounded elements – deriving by rounding back of flattened cells or from cells which never flattened at all – which divide to give small colonies. Multiplication can continue smoothly, but more usually it proceeds in an irregular and limited way until around the end of the second week, when one or more of the colonies enters on more continuous division and produces large numbers of cells. With subculturing these can be maintained in multiplication for long periods. Cell lines can be established fairly easily from large-scale cultures, probably from this cell-type, and one of these has been maintained for over years.

While actively dividing, or if crowded or in sub-optimal conditions, the multiplicative cells tend to remain rounded; but they can quickly spread out to the flattened form where circumstances allow (Fig. 1D). Soon after entering on more continuous multiplication, the flattened forms often fold into rigid-seeming, elongated, leaf-like forms; and at the same time the macrophage-like cells generally show an increased tendency to melanize, sometimes doing so en masse. As with the loss of fat and disruption of fat-body cells during their integration phase, these effects are probably degenerative, but may indicate that the cells have progressed to a higher stage of development.

The transformation of multiplicative forms to a definitely advanced product takes place in the third week. A variable number of cells, generally rather small, change over a period of three to four days to give a derivative which is larger (up to 2 times the original spread diameter), more irregular in outline, with the surface lined with fine foldings or striations, and normally showing an accumulation of fine oil-like droplets in the central cytoplasm (Fig. 1E). These cells (which are not the same as the earlier mentioned giant products) are much more changeable in shape than the previous circular cells, and can adopt a variety of different forms, but most eventually assume an elongate form, and tend to pull together into loose knots or to envelope other tissues. They appear to have lost the ability to divide and the tendency to melanize, and usually last only a week or so in the cultures, then degenerating and breaking up.

In the earlier paper it was remarked that the macrophage-like cells resembled Drosophila haemocytes, and their identification with the latter cell type now seems more certain. Larval haemocytes (and lymphocytes), when mounted in the culture medium, assume a form very similar to the normal circular forms, and show the same tendency to melanize; and during pupation haemocytes transform to a product which is like the larger derivative. In the larva, the haemocytes are mostly present as plasmatocytes, and the transformation at pupation is to a lamellocyte stage (Rizki, 1957). These are probably the stages represented by the culture elements.

Chitin-secreting cells

The most striking of the new cell types are the chitin-secreting cells. These become readily recognizable only in the second week of culture, when they form their chitin, but actually appear much earlier than this. Although difficult to detect, they can be identified within the first 24 h of culture, as small, round, rather clear cells occurring singly or in very tight clusters of up to about 20, distinguished only by the possession of a small refractile body or bunch of bodies (Fig. 2A, B). This body usually occupies a hollow in the centre of the cell or cluster: it appears to be associated with the chitin-secreting surface of the cells, and the hollow may result from the surface turning in on itself.

Fig. 2

(A) Early chitin-secreting cell from 2-day culture. (B) Cluster of about 8 early chitin-secreting cells from 2-day culture. (C) Cluster of older chitin-secreting cells, from 12-day culture, which has bloated out and secreted a layer of chitin. (D) Cluster from a 14-day culture which has secreted two successive layers of chitin, both with a pair of hairs at the upper end. (r, Refractile body; n, cell nucleus; ch, chitin layer.)

Fig. 2

(A) Early chitin-secreting cell from 2-day culture. (B) Cluster of about 8 early chitin-secreting cells from 2-day culture. (C) Cluster of older chitin-secreting cells, from 12-day culture, which has bloated out and secreted a layer of chitin. (D) Cluster from a 14-day culture which has secreted two successive layers of chitin, both with a pair of hairs at the upper end. (r, Refractile body; n, cell nucleus; ch, chitin layer.)

Over the first three or four days the cells show little change, but in the next few days they enlarge, and the refractile body becomes more prominent. Then, over a day or so, the hollow bloats out and the chitin appears as a continuous layer on its inner surface (Fig. 2C). The first chitinized elements generally appear around the tenth day, and they increase rapidly in numbers over the next two to three days to reach a maximum by about the fourteenth day, when they may be very numerous. The chitin sometimes melanizes to an extent (the refractile body often melanizes more heavily), and it usually soon detaches to lie as a free capsule within the cell or cluster. A new layer of chitin is then often formed outside the first (the process is thus akin to ecdysis), and a concentric series of capsules may be built up in this way. There is normally only one initial capsule per cell or cluster, but some clusters produce two or more, side by side, in association with separate refractile bodies.

Most of the capsules are simple spheres, of clear, plain, rather rigid-seeming chitin; but some are more complex, and the chitin can be textured, is sometimes striated with ridges, or appears more flexible. In certain cases these differences may merely be due to upsets in the chitin-forming process, but in others it seems more likely that they indicate activity of different types of chitin-secreting cells. Well-formed hairs are sometimes seen (which are duplicated when the capsule is duplicated, indicating that the doubling process involves a true replication) (Fig. 2D), and very occasionally tooth-like structures which could be larval mouth-parts or perhaps other larval structures. Thus, epidermis and tissue of the mouth-parts are probably among the types represented, and tissue of the cephalo-pharyngeal apparatus and the fore- and hind-guts may also be.

The chitin-secreting cells are a sensitive cell type, and their development is particularly subject to delay or inhibition, especially in the chitin-secretion phase. Markedly abnormal development can also occur; for example, some cells bloat out but show no obvious chitin-formation, and may go on swelling to give large, thin-walled bladders.

Tracheal cells

The tracheal cells, while also chitin-secreting cells, are distinct from the previous cell type. They develop later than most of the other identified elements, and appear more rarely, only arising in a small proportion of the cultures. They are usually first seen at about the beginning of the third week, appearing as rather small, clear cells that occur in very compact monolayer sheets and tend, at the edges of the sheets, to put out slender, tapered processes. Provided conditions are suitable, the cells divide rapidly, continuing to adhere tightly together. The sheets thus enlarge, and usually soon begin to extend tongues of tissue; and at a variable time after their appearance (generally within a day or so) the tongues roll up to form monolayer tubes (Fig. 3A, B). Whether the rolling marks a stage in the differentiation of the cells is not known, but it is probably intrinsic to the cells themselves rather than induced by a change in the culture as a whole, since it takes place at different times for different tongues on the same sheet. As with the enclosure of the chitin-secreting cells, it may result from a tendency of the chitin-secreting surface to turn in on itself.

Fig. 3

(A) Sheet of tracheal cells from 21-day culture. Has extended tongues of tissue, some of which have rolled up to form tubes. (B) Extensive tracheal tube system from a 28-day culture. (C) High-power view of the tip of a growing tracheal tube. (D) Imaginal disc cell vesicle from a 14-day culture. Part of upper surface only in focus.

Fig. 3

(A) Sheet of tracheal cells from 21-day culture. Has extended tongues of tissue, some of which have rolled up to form tubes. (B) Extensive tracheal tube system from a 28-day culture. (C) High-power view of the tip of a growing tracheal tube. (D) Imaginal disc cell vesicle from a 14-day culture. Part of upper surface only in focus.

The ends of the tubes normally remain unrolled in a flattened fan of varying size (Fig. 3C), and the tubes go on lengthening by continued cell-multiplication here and in the tube wall. Branching can occur, mainly by forking at the fans; the cells of the tube walls sometimes flatten and stretch; and very extensive systems can result. Occasionally the tubes produce an inner chitinous lining over a period of two to three days at an indefinite time after their formation; but when they do this they degenerate extensively and are lost. Unchitinized tubes, which usually stop growing after a few days, can persist unchanged for some time, but they tend to contract back to rounded masses and to detach, and also be lost.

The development often proceeds irregularly, with cell-multiplication and sheet or tube extension halting and then resuming; and quite frequently the tissue contracts back and deteriorates before it has got very far. In addition, the cell-sheets are often quite far advanced when first seen (not infrequently already in tube form), and it is probable that they can develop as condensed masses under certain circumstances, and only flatten out to the sheet form when conditions allow. This may partly account for the low frequency of appearance of the sheets in the cultures, although that, and the irregularity of development, are probably mainly due to the poor growth conditions by the third week when the cultures are heavily overgrown by other cell types. The flattening seems to be helped by medium changes (probably because of the proteolytic activity in the medium) and appearance of sheets and tubes can be increased by giving a 75% rather than a 50% medium-replacement to well-developed two- to three-week- old cultures.

Imaginal disc cells

The macrophage-like cells and the tracheal cells are only two of a number of multiplying cell types that appear in the cultures, but of the others only one has yet been identified; the imaginal disc cells. These are again small, clear cells, which normally occur in hollow, monolayer spheres or vesicles (Fig. 3D). A few such vesicles arise within the first 24 h of culture, but mostly soon disappear, and growing vesicles usually do not begin to appear until towards the end of the first week. They then increase in numbers over the next week or so, and can become very numerous.

Initially the vesicles are small, consisting of only a few cells (up to about 20), which are thinly stretched and tightly bound together – perhaps with cytoplasmic connexions, since the cell boundaries are usually not clear, and the cells are often pulled sharply out of shape and have cytoplasmic strands across the lumen of the vesicle. With cell multiplication, which proceeds from the start, they enlarge and the cells become less-stretched and more closely packed together. Growth is irregular, and usually slow to begin with, but becomes better towards the end of the second week. In the third week the vesicles can reach very large sizes, with many hundreds of cells. They tend to stop growing at the end of the third week, due to the overcrowding in the cultures, but if tissue is transferred to larger culture vessels they can be kept for several weeks longer, and may continue to enlarge, sometimes attaining several thousand cells. So far it has not been possible to establish permanently-growing cell lines from them.

The vesicles normally remain spheroid and monolayer while growing, but some become more irregularly shaped and they can develop local thickenings. The older, larger vesicles sometimes become folded and multilayered, and may contract into more-solid masses generally. In the third or fourth week of culture, some produce an inner detached membrane, possibly chitinous, the significance of which is not clear.

The tissue was suspected to be imaginal disc material from its similarity to the vesicles found by Schneider (1972) to develop from cut halves of late embryos cultured in vitro, and proved by her to be imaginal disc material by being shown to differentiate into adult fly structures when implanted into larvae to undergo metamorphosis. The same test was therefore applied to well-developed vesicles from our cultures, and again recognizably adult structures – sheets of chitin with various patterns of hairs and bristles – were obtained from transplanted vesicles and solid imaginal cell aggregates. This work will be reported on in greater detail in another paper (Dübendorfer, Shields & Sang, 1975).

Unidentified cell types

A number of other cell types appear in the cultures, but remain unidentified so far and will not be described in detail. They arise at various times-mostly in the second and third week – and with various degrees of probability, some occurring regularly and others only being seen rarely. They differ quite widely in size, form, and arrangement, but can be broadly classified as either fibroblast or epithelial type elements, with some of the latter tending to a spread macrophage-like form. A number are multiplicative, and can become very numerous; others have not been seen to divide, and remain few in number. The development of most has not been followed closely, and it is possible that some are different developmental stages of a single cell type. It may also be that some are unusual or aberrant forms of other products, though this only seems likely in one or two cases. They include the fibroblast-type cells reported in our earlier paper, which are the first, commonest, and most prolific of the multiplicative types to arise (appearing at around the seventh day, and multiplying – irregularly at first, but then more steadily – to give extensive networks and multilayer bundles by the third week), and these are shown as an example of the fibroblast class in general (Fig. 4A). One of the commoner multiplicative epithelial types is also shown, as a representative of the epithelial class (Fig. 4B).

Fig. 4

(A) Unidentified fibroblasts from a 21-day culture. (B) Unidentified epithelial cells from a 21-day culture.

Fig. 4

(A) Unidentified fibroblasts from a 21-day culture. (B) Unidentified epithelial cells from a 21-day culture.

By the third week the cultures have become very dense as a result of the growth and multiplication of the various cell types, and during this week there is usually a marked aggregation of the tissues into large, irregular masses. The cultures can be maintained for some time after this, and may show outbursts of growth by one or other of the multiplicative cell types – the imaginal disc vesicles, for example, reach their largest size in this later period – but there is little further significant new development, and a progressive decline and loss of most of the tissues takes place.

Our studies show that in cultures of dissociated gastrula-stage embryonic cells of Drosophila melanogaster a wide variety of different cell types develop, including most of the major tissue types of the emergent larva, and there is some organization of some of them into higher structures. That is, the cells can differentiate without being associated with tissues or organs. The differentiating cells appear to fall into two main groups, which broadly relate to two main pnases of development. A first group involves cells that differentiate very quickly into their final histological form, and mostly show only secondary, maturational and organizational, changes after that; a second group involves cells that remain quiescent for a week or more, then resume multiplication and only slowly differentiate into their final form (or do not achieve full differentiation in some cases). The first group includes nerve, muscle, fat-body, chitinsecreting, and a first stage of development of the macrophage-like cells; the second group includes a second stage of development of the macrophage-like cells and tracheal, imaginal disc, and a number of unidentified fibroblastic and epithelial cells. A proportion of the nerve and muscle cells go through an early, limited phase of rapid division in the first few hours after establishment of the cultures, and a similar phenomenon seems likely for some – perhaps all – of the other cell types.

The identification of the cell types rests mainly on visual evidence, but is reasonably certain in most cases. The nerve, muscle, chitin-secreting, and tracheal cells are so distinctive in their form, organization, and activities as to require little further confirmation. The fat-body cells, when well developed, are also characteristic enough in appearance to make their identity fairly unquestionable, though other cell types can also accumulate refractile fat-like droplets to some extent, and may be included among the lesser fat-filled cells in the cultures. With regard to the imaginal disc cells, differentiation of vesicles to adult structures on implantation into larvae and passage through metamorphosis is a decisive test of their character. But it has not been definitely established that all vesicles so differentiate, or that all differentiated exoskeletal structures arise from vesicles. It is possible that some vesicles are of a cell type other than imaginal disc, and that disc cells may sometimes develop in a non-vesicular way. Distinctly non-imaginal disc looking vesicles do appear in the cultures; and there is some evidence that, as with the tracheal tissue, imaginal disc cells can develop in contracted masses under certain conditions.

The main doubt, however, is as to whether the macrophage-like cells are haemocytes, and further data are needed before they can be confidently so designated. There is evidence that they are phagocytic, both in the early and the transformed stage, and this would support a haemocyte identification; but it is likely that other cells also have this property (e.g. yolk cells). It may be that the early appearing macrophage-like cells are, in fact, heterogeneous, since they show a wide variability of form.

Schneider’s (1972) prior observation of development of imaginal disc vesicles from cultured fragments of late embryos has already been noted, and in the same cultures she also saw multiplying macrophage-like cells. She was able to establish cell lines from her cultures which she believed to derive from the vesicles, but one of the lines (line 1) appears much like our cell lines, which we think derive from the macrophage-like cells, and we should note that in our experience the latter cells tend to rapidly outgrow other cell types. With our lines, as with the Schneider line 1, there was an early change from presence of a high proportion of normal macrophage-like cells to a predominance of small rounded cells, apparently due to selection out of faster-growing elements. The ease with which we can obtain cell lines from gastrula-stage embryos does not support Schneider’s suggestion that late embryos may be better than early ones for deriving lines.

Fat-body, tracheal, and chitin-secreting cells have not been reported for any Drosophila embryo cultures before; nor do they appear to have been positively observed for any other insect embryo cell culture. Mitsuhashi & Maramorosch (1964) noted emergence of two cell types that accumulated refractile droplets in cultures from Leafhopper embryos, and suggested that one of them might derive from fat-body, but this was a wandering (macrophage-like) cell type and, in fact, the other, epithelial type, looks more like our fat-body cells, though in an early stage of development. Presence of tracheoblasts in organ cultures of the nervous system and oesophagus of young cockroach embryos has been claimed (Chen & Levi-Montalcini, 1969; Aloe & Levi-Montalcini, 1972), but the evidence is slight. Chitin-secreting cells in the form we report have nowhere been described but, interestingly, deposition of cuticular material by vesicles that seem similar to our imaginal disc vesicles has been reported in cockroach leg regenerate (Marks & Reinecke, 1964; Marks & Leopold, 1971), cockroach embryo organ (Larsen, 1967), and mosquito halved-embryo cultures (Peleg & Shahar, 1972). In the first case, at least, the cells were almost certainly epidermal. This heightens our caution regarding the nature of vesicles in our cultures

Some mention should be made of important larval cell types which have not been recognized as occurring in our cultures. They include mid-gut (and perhaps the rest of the digestive tract) gonadal, Malpighian tubule, and salivary gland cells. Some or all of these could be among the unidentified cell types which appear, but if so, criteria other than cell form and arrangement will probably have to be used to identify them. One such test might be the status of the chromosomes. Aceto-orcein squash preparations of our cultures reveal the presence of cells with polytene chromosomes, and there appears to be variation in the degree of polytenization. Some of the polytene chromosomes are probably associated with the fat-body cells, but the more striking ones seem likely to involve other cell types.

Seecof, Alléaume, Teplitz & Gerson (1971) reported the presence of glial cells in association with the nerve cells in their Drosophila embryonic cultures; and Chen & Levi-Montalcini (1969) made a similar observation for cockroach embryo brain cultures. In our cultures the nerve axons are always naked. The cells described by Chen & Levi-Montalcini seem rather similar to the much-flattened forms of our early macrophage-like cells; but a stronger proof of identity than morphology is required before these can be claimed as glial cells.

The Drosophila egg is of the mosaic type, determination taking place in the blastula stage, when cells are first formed (Poulson, 1950), and it is likely that with -old embryos the cells are well-established as to their ultimate character. The pattern of development of our first group cells is therefore understandable: these are probably cells that are rigidly programmed to differentiate to their specific form, and already well advanced along the differentiation pathway at the time of going into culture. They are able to continue developing at much the same rate as they would in the whole organism. Further development of the second group cells apparently entails continued multiplication, and it seems that this is quickly inhibited in the in vitro situation. The multiplication only resumes after a delay – perhaps through adaptation of the cells, or a general conditioning of the culture medium, or the supply of specific factors by one or more of the other developing cell types – and then proceeds slowly, so that the whole further programme of development is slowed down. The ultimate differentiation of the cells might also be dependant on supply of specific factors by other cells in some cases, and so again subject to delay or non-occurrence. It is worth noting that the resumed multiplication and differentiation follow on maturation of the fat-body cells, which would be expected to have a beneficial conditioning effect on the medium and perhaps also to produce specific growth and developmental factors. Formation of chitin by the chitin-secreting cells of the first group also occurs after maturation of the fat-body cells, perhaps for the same reason.

This work was supported by grants from the Science Research Council of Great Britain. A.D. was in receipt of a Royal Society European Exchange Fellowship.

Aloe
,
L.
&
Levi-Montalcini
,
R.
(
1972
).
Interrelation and dynamic activity of visceral muscle and nerve cells from insect embryos in long-term cultures
.
J. Neurobiol
.
3
,
3
23
.
Chen
,
J. S.
&
Levi-Montalcini
,
R.
(
1969
).
Axonal outgrowths and cell migration in vitro from nervous system of cockroach embryos
.
Science, N.Y
.
166
,
631
632
.
Dübendorfer
,
A.
,
Shields
,
G.
&
Sang
,
J. H.
(
1975
).
Development and differentiation of imaginal disc cells from early Drosophila embryos cultured in vitro
.
J. Embryol. exp. Morph. (In the Press)
.
Larsen
,
W.
(
1967
).
Growth in an insect organ culture
.
J. Insect Physiol
.
13
,
613
619
.
Marks
,
E. P.
&
Leopold
,
R. A.
(
1971
).
Deposition of cuticular substances in vitro by leg regenerates from the cockroach Leucophaea maderae
.
Biol. Bull. mar. biol. Lab., Woods Hole
140
,
73
83
.
Marks
,
E. P.
&
Reinecke
,
J.
(
1964
).
Regenerating tissues from the cockroach leg
.
Science, N.Y
.
143
,
961
963
.
Mitsuhashi
,
J.
&
Maramorosch
,
K.
(
1964
).
Leaf-hopper tissue culture
.
Contr. Boyce Thompson Inst. Pl. Res
.
22
,
435
460
.
Peleg
,
J.
&
Shahar
,
A.
(
1972
).
Morphology and behaviour of cultured Aedes aegypti mosquito cells
.
Tissue & Cell
4
,
55
62
.
Poulson
,
D. F.
(
1950
).
Histogenesis, organogenesis, and differentiation in the embryo of Drosophila melanogaster Meigen
.
In Biology of Drosophila
(ed.
M.
Demerec
) (pp.
168
247
).
New York
:
Wiley
.
Rizki
,
M. T. M.
(
1957
).
Alterations in the haemocyte population of Drosophila melanogaster
.
J. Morph
.
100
,
437
458
.
Schneider
,
I.
(
1972
).
Cell lines derived from late embryonic stages of Drosophila melanogaster
.
J. Embryol. exp. Morph
.
27
,
353
365
.
Seecof
,
R. L.
,
Alléaume
,
N.
,
Teplitz
,
R. L.
&
Gerson
,
I.
(
1971
).
Differentiation of neurons and myocytes in cell cultures made from Drosophila gastrulae
.
Expl Cell Res
.
69
,
161
173
.
Seecof
,
R. L.
,
Donady
,
J. J.
&
Teplitz
,
R. L.
(
1973A
).
Differentiation of Drosophila neuroblasts to form ganglion-like clusters of neurons in vitro
.
Cell Differentiation
2
,
143
149
.
Seecof
,
R. L.
,
Gerson
,
I.
,
Donady
,
J. J.
&
Teplitz
,
R. L.
(
1973b
).
Drosophila myogenesis in vitro: The genesis of ‘small’ myocytes and myotubules
.
Devi Biol
.
35
,
250
261
.
Shields
,
G.
&
Sang
,
J. H.
(
1970
).
Characteristics of five cell types appearing during in vitro culture of embryonic material from Drosophila melanogaster
.
J. Embryol. exp. Morph
.
23
,
53
69
.