ABSTRACT
By placing blood plasma coagula containing growing cultures (fibroblasts from chick embryo) in a vertical position (vertical cultures) certain physical and quantitative chemical alterations of the medium occur under the influence of gravity; the effects of these upon the growth, morphology and products of differentiation (fibre formation) of the cultures are described.
In vertical cultures the normal circular shape of the cell colony remains undisturbed, indicating that neither gravity nor a gradient of metabolites seems to play a role controlling the direction of movement of fibroblastic cells.
Vertical cover-slip cultures show a greater radial extension (outgrowth) than the horizontal controls, apparently due to increased cell migration rather than to augmented cell division.
Vertical cultures of fibroblasts, both on cover-slips and in flasks, do not show the usual appearance of visible fat droplets. If fat appears at all in vertical cultures a fat-free sector can sometimes be observed in the lower half of the culture. Flattening of the coagulum retards but does not prevent the appearance of visible fat in normal cultures.
Under certain conditions the appearance of fat is an indicator of reduced activity since activity of fibroblastic cells (growth by migration) can be increased by preventing the accumulation of fat droplets in them. The latent period of offspring of a 10-day-old fat-free vertical culture is shortened as compared with that of offspring of the very fatty horizontal control of the same age. The greater activity observed in the vertical cover-slip cultures is apparently due to the non-appearance of fat.
Vertical cover-slip cultures after three days’ cultivation show a slightly stronger formation of argyrophil fibres in the outer growth zone than the controls, but culture pairs fixed at a later time show no appreciable difference in fibre formation.
INTRODUCTION
For the majority of experiments on tissues growing in vitro it is necessary to use a supporting structure for the cells, and up to the present time no medium has been found to be as suitable as blood plasma which is allowed to coagulate around the tissues. But not only does the coagulum introduce unknown variables into the physico-chemical environment of the tissue cells, it also exerts mechanical effects of fundamental importance. The experiments described in this paper were undertaken with a view to finding out some of the essential physical characteristics of the coagulum and to investigate their importance in regard to the behaviour of cells growing in it. The plasma for tissue cultures is generally avian, and such (fowl plasma) was used during these experiments. The intimate structure of the coagulum is still a matter for conjecture, e.g. the fibrin present in it is ultra-visible and not stainable by Weigert’s method (Mayer, 1933), and although the coagulum contains a considerable quantity of fluid it is uncertain how this is held, and what relation it bears to the more solid fibrin. A process, intimately connected with this relationship, is the so-called “liquefaction” of the coagulum in which process the cells almost certainly play a great part (Fischer, 1929; Mayer, 1930). Embryo extract (Losee & Ebeling, 1914; Demuth & von Riesen, 1928; Willmer & Kendal, 1932) and even the enzymes of plasma itself (Demuth & von Riesen, 1928) are also said to be contributing factors. For, as the coagulum of the tissue culture is not simply clotted plasma, but is often composed of a mixture of plasma and embryo extract or Tyrode solution, the appearance of fluid may not be comparable with the exudation of serum from a blood clot. Certain experiments therefore were designed to investigate the appearance of fluid in the coagulum of the tissue culture. By placing the coagulum in a vertical position, in which the effect of gravity is of course maximal, instead of in the usual horizontal position, any fluid which is formed can drain away and so give an indication of its rate of formation. This procedure was performed on cover-slip coagula (16 mm. in diameter) and on coagula in Carrel flasks (diameter 35 mm.), both with and without the presence of living cells (“vertical experiments”).
In the course of these experiments it became evident that the “vertical method” produces changes in the coagulum, which are clearly reflected in the behaviour of the cell colony implanted in it. Thus, for instance, it is possible by using this method to obtain cultures which remain entirely free from fat droplets. As, therefore, the facts emerging from these investigations have a significance beyond the restricted field of the special experimental conditions, the experiments on cell colonies in vertical coagula will be described more in detail after the experiments performed on the coagulum itself have been briefly reported.
VERTICAL COAGULA EXPERIMENTS
The embryo extract used for making certain of the coagula was prepared as described by A. Fischer (1930). Seven-day-old chick embryos were ground to a pulp, the pulp was centrifuged and the supernatant fluid pipetted off (first centrifugate). The sediment was then thoroughly mixed with an equal quantity of Tyrode solution and centrifuged again. The supernatant fluid so obtained was the second centrifugate. A repetition of this operation produced the third centrifugate. For the cover-slip coagula—according to the usual technique—a mixture was used consisting of the second and third centrifugates in equal parts. For the flask coagula the first centrifugate was diluted to 10−20 per cent, with Tyrode solution. The coagula were incubated at 39° C.
From vertical cover-slip coagula (one drop plasma + one drop embryo extract on mica slips), but far more regularly from vertical flask coagula of various compositions, there occurred a great issue of fluid, leading to a consolidation and flattening of the coagulum. This flattening was asymmetrical, above more than below (Text-figs. 1 and 2). Measurements of the thickness of the coagulum revealed that the original thickness, which was about 0·8 mm. in the middle, decreased after 48 hours to about 0·2 mm., afterwards decreasing still further. The issue of fluid from coagula of plasma and Tyrode solution was the same as from plasma and embryo extract, indicating that the exudation of fluid is purely a mechanical process and not due to a chemical effect (fibrinolysis) on the part of the embryo extract. Also proteolytic enzymes of the plasma itself cannot be responsible for the issue of fluid, for, if vertical coagula of undiluted plasma were produced, there was no fluid issue in coverslip preparations, and fluid only appeared very slowly in flasks. Text-fig. 3 shows graphically the rate of issue of fluid from vertical flask coagula of various compositions. It is interesting to note that the issue of fluid for a given plasma concentration was quicker and greater from a coagulum formed by spontaneous clotting than from one with extract. The controls, i.e. coagula in the usual horizontal position, never showed the least appearance of fluid.
When fibroblast cultures were grown in vertical coagula, there was also regularly produced, both in cover-slip preparations and in flasks, a great issue of fluid, leading again to an asymmetrical flattening of the coagulum. Vertical flask experiments, in which the fluid exuded can be measured, showed from coagula with two growing cultures a little more fluid issued than the vertical control coagula of similar composition without tissue. But the difference in most cases was only slight (an average of 1·25 c.c. as compared with 1·1 c.c. in a period of 10 days), indicating that living cells do not add significantly to the amount of fluid, if at all. They break down the fibrin structure, presumably by digestion, as can be shown by the depression or funnel (Mayer, 1933) in the surface of the coagulum in the immediate vicinity of the culture, a phenomenon which also regularly occurred in vertical cultures, but was never obtained in culture-free coagula, even if a non-living material had been put in the centre of the coagulum. The bearing of these experiments on the process of “liquefaction” will be discussed below.
VERTICAL CULTURES EXPERIMENTS
The changes introduced in the system coagulum-cell colony by the vertical method are: (1) gravity can exert its maximum effect, (2) consolidation of the clot occurs with a progressive flattening, (3) the withdrawal of a considerable amount of fluid from the coagulum involves the removal of food substances and of metabolites resulting in definite gradient systems throughout the clot.
Mechanical alterations of the medium are said to cause especially morphological changes in the cells (Fischer, 1930). Weiss (1929) suggested that the withdrawal of fluid from the coagulum was a factor concerned in producing narrow elongated cell shapes. Willmer (1933) found the cells in a shallow clot more transparent and less fatty than in a deeper coagulum. Again, Baitsell (1915), observing the formation of a fibrous tissue in vitro, assumed that a transformation of fibrin took place and that in this process a mere consolidation of the fibrin was of great importance. The vertical method, therefore, suggested itself as being suitable for experimentally testing such suggestions and hypotheses and thus amplifying the knowledge of the physiology of cells cultivated in vitro.
Material
Mesenchyme tissues were explanted from heart, perichondrium (femur) and bone (os frontale) of 10−12-day-old chick embryos. The strains so obtained (heart fibroblasts, perichondrial and periosteal fibroblasts) were only taken for experiments after they had been subcultured and transplanted at least seven times since removal from the embryo, thus representing so-called “pure cultures”. By cutting one culture into two equal halves, experimental and control tissues of the greatest similarity were obtained; throughout the work such sister cultures are designated a and b. Experiments were performed both by the cover-slip and the flask method.
The embryo extract was prepared as described above. For the cover-slip coagula, to one drop of hen plasma was added either one drop of the second centrifugate (for heart fibroblasts) or one drop of a mixture of the second and third centrifugates in equal parts (for perichondrial and periosteal fibroblasts). For the flask coagula, the first centrifugate was diluted to 10−20 per cent, with Tyrode solution and the composition of the coagulum was generally 0·5 c.c. plasma + 1·0 c.c. extract solution.
Technique
Cover-slip cultures
On mica slips circular coagula of equal size (diameter about 16 mm.) were spread. The implanted cultures were brought exactly into the middle. The outside of each slip was marked with ink at two points so that the original central point of the tissue could be fixed when the surface measurement was obtained later. After clotting was complete, the slides containing the experimental cultures were placed vertically. The controls were kept in horizontally hanging coagula as usual.
Flask cultures
Two cultures were put in each flask so that in the vertical position one was in the upper part of the coagulum and the other in the lower part. The fluid that issued from the coagula of the vertical flasks was pipetted off every other day. During the course of the experiment the cultures were neither washed nor fed.
In the tables, figures, etc., the vertical experiments are, when necessary, marked with an arrow pointing downwards (↓). The experimental arrangement will be obvious from Text-figs. 1 and 2.
Measurements of growth were taken every 24 hours by the surface measurement method of Ebeling (1921). This measurement—giving, strictly speaking, only the “Wachstumsbilanz” of the cell colony (Mayer, 1930)—is sufficiently accurate and allows growth comparisons between culture pairs, provided that these correspond in cell density, which can be approximately estimated by careful microscopic examinations. The surface areas of halved cultures vary on the average by less than 10 per cent., provided that the surfaces of the culture pairs after 24 hours are in good agreement with one another.
Growth of the vertical cultures
Cover-slip cultures
The first thing to be noted is that the normal circular culture shape, being the resultant of the growth intensity in different directions, was unaltered in the vertical experiments. The horizontal and vertical radii of each culture were measured, but no significant difference was found. It was clear also that the whole vertical cell colony did not change its position. These results not only negative any direct effect of gravity upon the cells, as suggested by Uhlenhuth (1915), but they also show that the various gradient systems caused by gravity in the vertical coagula are equally ineffective in producing any disturbance in the circular shape of the cell colony. The vertical cover-slip cultures seemed, if anything, to approach to the ideal circle rather more often than their horizontal controls. The explanation of this was found by measuring the extent of growth of vertical coverslip cultures and their controls. Contrary to what might have been expected since food substances are removed in the fluid that exudes from vertical coagula, the vertical cultures attained even larger sizes than their controls. During the first 3 days it is quite justifiable to compare the areas of the culture pairs as a measure of growth, since during this time the cell density of the vertical cultures and their controls differed but little (Plate I, figs. 1 and 2). When the experiment was more prolonged, however, the cell density in some of the vertical cultures became less. Therefore the final results expressed as surface measurements give sometimes a comparison of the radial extension of the colonies rather than of their true nett growth. Table I shows that on the average the final size reached by the vertical cultures is 20 per cent, in excess of that of the controls, the average ratio exper. /control being 1·2 after 160 hours. The calculation of a mean value seems permissible, as the deviations are all in the same direction. In isolated cases the ratio even reaches a value as high as 1·5. In some of the cases in which the quotient is only 1, the unequal initial sizes of the cultures should be taken into account, the controls being still greater than the experimental ones after 24 hours (Nos. 1 and 7 of Table I). In such cases the experimental cultures either caught up or overtook the controls. The increase in the final size of vertical cultures was due to a longer period of growth, a greater growth rate, or to both of these (Text-figs. 4, 5 and 6). The difference in growth rate between experiment and control became evident after 3 days, at about which time the growth of the controls tends to subside.
In order to estimate the importance of cell multiplication, as distinct from increase in colony size, six cover-slip culture pairs (two at a time) were fixed (Bouin’s fluid) and stained (acid haemalum) at different intervals between the first and third day after transplantation. The mitoses were counted throughout the whole culture by systematically covering the field with in. oil immersion lens and squared eyepiece. The counting is naturally most difficult in the central region of the culture where the cells are lying very densely packed, but in the subcultured tissues used in these experiments it was found to be quite possible. Since the culture pairs were comparable in origin, age, cell density and size, the figures of mitoses so obtained gave a sufficient answer to the question whether the factor of cell multiplication was influenced by the change of the medium. The results are given in Table II, which also gives the age and surface areas of the cultures when fixed. The table shows clearly that at these given times there was no difference in number of mitoses between the vertical cultures and their controls. This indicates that the stronger radial extension of the vertical cover-slip cultures is probably due to migration rather than to an increased rate of mitoses. A further fact of importance which emerges is that, in spite of a great withdrawal of fluid, food substances, etc., from the medium, the rate of cell multiplication was not affected.
Flask culture
In contrast to cover-slip cultures, vertical flask cultures did not show a greater radial outgrowth than the horizontal controls. In most cases they reached the same size (Text-fig. 7). A possible explanation may depend on the fact that in flask cultures the growth period is longer than in cover-slip cultures, and at the time when the vertical cultures in flasks might prove their further capabilities, both they and their controls have already reached the end of their growth periods as determined by the population density, amount of food substances present, etc., with the result that any favourable condition—as apparently given through the vertical position (see the morphological behaviour)—is without effect in producing a further increase in growth. When two sister cultures were placed in each vertical flask, so that the a culture grew in the upper part of the coagulum, the b culture in the lower, always somewhat thicker part, a similar result was obtained, the growth curves of both cultures showing almost ideal concordance. Both series of flask experiments seem to indicate that the greater radial outgrowth observed in the vertical cover-slip cultures was not a direct and simple consequence of the flattening of the coagulum, but depends on other factors also. This becomes evident from a consideration of the morphological characteristics of the cultures.
Cell morphology of the vertical cultures
Contrary to what might have been expected from the work of Uhlenhuth (1915) and Weiss (1929), the cells in vertical cultures were not markedly elongated or narrow but showed no significant difference in shape from those in horizontal cultures. These results are interesting as indicating that the fluid content in the coagulum, once it is formed, has very little influence on the shape of fibroblasts growing in it.
On the other hand, there is a noticeable difference in the number of fat droplets visible in the vertical cultures as compared with those in horizontal controls. In the latter from the third day after transplantation cultures regularly show an increasing quantity of fat droplets within their cells, while in the vertical cultures either no fat appeared at all or only a minimal quantity confined to the peripheral cells. The results of thirty-three cover-slip experiments, some of which were carried on for as long as 16 days, are given in Text-fig. 8. Since no quantitative method is available for determining the visible fat, the amount present was estimated by direct observation and recorded by “+” and “−” signs. In twenty experiments the vertical cultures remained quite free from fat during the period of observation, while in nine experiments there was a very slight appearance of fat but only after the seventh day. The greatest fat content which was observed in the vertical cultures—and that only in four cases—was signified with one “+”, and this occurred on the eighth day at the earliest. In the controls, on the other hand, fat was present from the third day in increasing amounts. Besides this arbitrary method of estimating the fat, in several experiments photographic records were taken at intervals during the life of the cultures. From the series obtained, typical examples are given. Plate I, figs. 3 and 4, show, experimental culture and control on the seventh day after transplantation. In strong contrast to those of the control whose cytoplasm is loaded with fat globules, the cells of the vertical culture are seen to be transparent, quite clear and almost fat-free. Plate II, figs. 1 and 2, show two complete cultures. Preparations fixed and stained with Sudan III confirmed these observations. When any fat did appear within the cells of vertical cultures it was found that in the lower half of the culture there was often a sector of variable size which was relatively or quite free from fat. Such a sector is shown on Plate II, fig. 1, in the lower part of the culture.
Vertical flask cultures showed nearly the same fat-free behaviour as the coverslip experiments (Text-fig. 9). From sixteen experimental cultures only six showed, after the sixth day, a visible fat content which could be signified with one “+”. The rest remained almost free of fat droplets. In some cases again a fat-free sector appeared in the lower half of the culture. Though no explanation can be given for this phenomenon, the result obtained from vertical flasks containing two cultures, one in the upper part of the coagulum, the other in the lower part, each showing the fat-free sector and an equally small content of fat in the upper part, indicates that the cause of the appearance of this sector is connected with a gradient in the cell colony itself rather than with anything in the medium.
Although fibroblasts derived from heart have a stronger tendency to accumulate fat when growing under normal conditions than have periosteal or perichondrial fibroblasts, the different fibroblastic strains used in these experiments all behaved in the same way, in that they all remained free, or almost free, from visible fat when growing in vertical coagula.
Depth of the coagulum and fat appearance
Willmer (1933) found that the cells of a tissue growing in a shallow clot were “relatively clear and free from fat” (time of observation 4 days), and that the distance from the surface was responsible for the type of growth. Therefore special investigations—in particular over a longer period and using smaller depths—were necessary, to find out whether the flattening of the vertical coagula alone might give a sufficient explanation of the non-appearance of fat in the vertical cultures.
Coagula (one drop plasma + one drop embryo extract) were spread out on mica slips so that their diameter was about 24 mm. In this case the thickness of the coagulum (measured microscopically) is about 0·2 mm., which approaches very closely to the thickness of the flattened coagulum in vertical cultures. Controls were cultivated in the usual coagula about 16 mm. in diameter and with an initial thickness of 0·8−1·0 mm., and some were kept hanging horizontally, others as vertical cultures. Thus the amount of medium and the concentration of the plasma were kept constant, and only the depths of the coagula were varied within the required limits. The result was: the cultures in very shallow horizontal clots became increasingly fatty from the fifth day onwards in contrast to the vertical cultures which remained again almost entirely free from fat (Plate III, figs. 1 and 2). But when compared with the usual horizontal controls with thicker coagula, the appearance of the fat was somewhat delayed and on the whole less marked. From this it follows that the flattening of the coagulum most probably does have an inhibitory action on the process of fat appearance in vertical cultures, but it is certainly not the only factor concerned.
The amount of visible fat and latent period
As the cells in vertical cultures always looked very transparent and healthy even if kept for much longer time in the culture chamber than is usually done, the question arose : Are they actually in a more healthy state or not? This could be examined by subculturing and transplanting them into a new medium. A culture when trans planted usually shows a latent period of varying duration. It is especially long, for example, when a cover-slip culture is subcultured after three or more days instead of after two, indicating that one cause of the lag phase probably lies in the chemical or physical state of the culture itself when it is placed in the new medium. The fatty state of the cells might well be correlated with the lengthening of the latent period. The fat-free condition of the vertical cultures therefore opened up a possibility of examining this point and deciding whether fat-free cells are relatively healthy and active or the reverse. Vertical cover-slip cultures and their controls, after having been kept for 10 days, were halved, transplanted and cultivated horizontally on mica slips. In this procedure it was found that the vertical cultures had become closely adherent to the mica as a thin film and were consequently difficult to transplant, and that the daughter cultures clumped readily in the Ringer bath. On the other hand, the daughter cultures of the control which were very fatty, were easy to transplant as broad discs. These cultures were derived from the two cultures shown on Plate I, figs. 3 and 4, and the results obtained are shown in Text-fig. 10 and Plate III, figs. 3 and 4. They show clearly that in the first 48 hours the velocity of the radial outgrowth of the offspring of vertical cultures was certainly greater than that of the controls. The absence of fat droplets may give this advantage. The cells of the transplanted vertical culture seem to be actually in a more reactive state, leading to a quicker migration and division, than the very fatty cells of the transplanted control which need a longer period for recovery. The final size of all the cultures is fairly small as a consequence of the unusual length of time (10 days) for which they were kept in the medium prior to the second part of the experiment. These experiments allow the conclusion to be drawn that one factor causing the usual lag phase is connected with that state of the cells which has its expression in the appearance of fat droplets. The experiments also show that the vertical cultures are perfectly healthy and can be transplanted successfully.
Formation of argyrophil fibres
Cells of pure mesenchyme cultures, when growing in a plasma coagulum, regularly build up an extracellular argyrophil fibre system corresponding to the reticulum fibres of the animal organism. The development of these fibres is due to the activity of the living cells, and the degree of fibre formation depends on their growth intensity (Bofill-Deulofeu, 1932). The part played by the medium itself in this process is not yet understood. Baitsell (1915), as mentioned, assumed that a mere consolidation of the fibrin present is responsible for the fibre formation. As the vertical method affords an ideal case of such a consolidation, an investigation of this point suggested itself.
Cover-slip culture pairs were fixed in formol at different times after implanting and treated according to the silver impregnation method of del Rio Hortega (see Bofill-Deulofeu, 1932). Plate III, fig. 5, shows the normal fibre formation in a culture of perichondrial fibroblasts. While several vertical cultures 3 days after transplantation showed a slightly stronger formation of fibres in the outer growth zone than the horizontal controls, others fixed and stained at a later period (fifth or seventh day) showed that eventually there was no marked difference in fibre development between vertical cultures and their controls. This favours the theory that fibre formation lies more or less entirely with the cells and is not greatly influenced by consolidation of the medium.
DISCUSSION
The results obtained from the vertical coagula experiments have an interesting bearing on the nature of the appearance of fluid in tissue cultures. It is evident from the ease with which gravity effects the separation of the fluid from the more solid part of the clot that the fluid in the plasma coagulum is held very loosely and a great portion is free rather than “bound”; this seems to give a substantial support to A. Fischer’s conception (1933) of the coagulum as a “Fibrinschwamm” in the pores of which—so to speak—the fluid is contained. This simple exudation of fluid is almost purely mechanical and is to be distinguished quite clearly from fibrinolysis. This latter process occurs when cells are present which, by destruction of the fibrin structure, liberate the enclosed fluid. This fluid may appear if conditions allow. This explains easily the phenomenon seen in hanging-drop cultures when the fibrin structure immediately round the cells is destroyed to a great extent and the fluid so liberated exudes and collects as a drop just below the culture. It also well accounts for the manner in which cell colonies often shrink up into a ring surrounding a cavity in the coagulum. This process is known as the “liquefaction” proper. Here almost complete fibrinolysis may occur in the vicinity of the cells, but whether or not this change of the coagulum becomes visible depends on the behaviour of the cell colony. Sometimes a small flaw occurs in the cell network which may gradually enlarge owing to the tension exerted by the cells, or which may, if that tension is great enough, suddenly spread to the whole culture so that nearly all the cells become quickly grouped together on one side of a large circular fluid-filled space which is surrounded on the outside only by a few cells which were originally the most peripheral cells of the colony. Syneresis on the part of the coagulum, as suggested by Grossfeld (1934), seems to play but a small part in this process, if at all. The tension in the cell net is the greater the faster the peripheral cells migrate outwards ; and as the rate of migration is increased with increasing concentrations of embryo extract (Willmer & Jacoby, 1936), “liquefaction”, i.e. bursting of the culture, will necessarily take place more frequently in higher concentrations of embryo extract. This is what actually happens (Parker, 1929). At the same time fibrinolysis seems to be speeded up, and probably both these processes, migration and fibrinolysis, are closely connected with one another.
From the results obtained on vertical cultures perhaps the most interesting point which arises concerns the appearance and non-appearance of fat in tissue culture cells. Broadly speaking, fat becomes visible in cells from two causes. It may be storage material or a sign of degeneration, and it may be difficult, if not impossible, to decide, in any one particular case, which is the immediate cause. Cells irradiated by ultra-violet light become fatty and degenerate (Mayer & Schreiber, 1934). Cells growing in high concentrations of embryo extract become fatty. Cells growing deep in a coagulum become fatty, while those growing in shallow coagula remain relatively free from fat (Willmer, 1933). And now it is found that the cells of the vertical cultures remain almost entirely free from fat, but perfectly healthy and, as transplantation experiments show, they can be subcultured and then grow even more readily than do those of the horizontal controls whose cells are laden with fat. It was shown that the flattening of the coagulum, probably by facilitating the oxidation processes within the cells, retards the appearance of fat, but does not prevent it from appearing. Thus the vertical culture experiments confirm to some extent, but also amplify, the findings of Willmer (1933).
Horizontal cultures in their stagnant medium get an excess of food substances and lay down storage fat, which generally seem to inhibit cell movement to some extent. In vertical coagula the draining away of fluid removes this excess and the cells remain “lean” and active. This would afford an explanation of the stronger radial outgrowth shown in the vertical cover-slip cultures as compared with the horizontal controls. This was only observable after the third day at which time fat distinctly appears in the controls. The fact that the fat-free cells of vertical cultures migrate extensively, can be subcultured readily, and show the same numbers of mitoses as those in horizontal cultures suggests that, in the latter, growth-promoting substances and food substances are probably present in excess. This excess of food substances actually inhibits the movements of the cells on account of the amount of fat which it causes to be laid down in the cells. The removal of metabolites which takes place in the vertical cultures may be another factor effecting the fat-free condition of the cells. Actually then, on these lines, the cells in vertical cultures are the more healthy. However, it should be pointed out that the invisibility of fat does not necessarily imply its absence. Normal heart muscle which looks to be free from fat is shown on chemical analysis to contain as much fat as the degenerated muscle the cells of which are laden with visible fat (Starling, 1933). This indicates that in the appearance of fat droplets in the cytoplasm the essential factor is often a change in the physical state of the protoplasm. Therefore it may be possible that the nonappearance of fat in vertical cultures may be due, in addition to other factors, such as the flattening of the clot, and a reduced uptake of fat from the medium, to a kind of stabilisation of the physical state of the cytoplasm, presumably as a consequence of the consolidation of the coagulum.
The second significant point in connection with these vertical cultures is that, in spite of the withdrawal of fluid from the coagulum in one direction, in spite of the action of gravity, and in spite of the consequent gradients of metabolites and the directional stresses set up in the medium, there is no sign of any abnormal orientation of the outgrowth, which remains perfectly regular and circular, nor do the cells composing it alter their morphology except in relation to their fat content. Moreover, the withdrawal of fluid and drying of the system whiçh occur have no effect on the rate of cell division, but it should be pointed out that this drying is perhaps different from the type of drying which is correlated with declining growth rate in developing organisms (Needham, 1931). The fact that the rate of cell division in these cultures is not affected by the strong withdrawal of fluid indicates that either there is a sufficient amount of growth-promoting substances still remaining in the coagulum, or the concentration present at the very beginning decides the whole course of growth.
Again, it was stated that in the vertical flasks there was no difference in surface growth between experimental and control cultures in spite of different depths of the coagula. This result seems to differ from certain flask experiments by Willmer (1933); the apparent discrepancy, however, will be strongly reduced, if one takes into account that in the experiments of this author the absolute amount of fibrin present had to be varied, whereas in the vertical flasks it was possible to decrease the thickness of the coagulum without altering the amount of fibrin. This would indicate that an important factor determining growth in surface is the absolute amount of fibrin present round the culture.
Finally, it may be noted that the withdrawal of fluid from the medium, such as occurred in these experiments, had no influence on the extent to which argyrophil fibres were developed in the medium. This is certainly an argument in favour of the hypothesis that the major part in fibre formation is played by the cells whose products act on the proteins in the medium, and, as long as the latter are not too much reduced in amount, fibre formation remains unaffected, except by changes in the cells themselves. Moreover, fibre formation like other differentiation processes seems to be to some extent independent of growth, for although horizontal cultures cease to grow before sister cultures in the vertical position, they continue to lay down fibres in their peripheral parts, so that eventually there is no difference in the distribution of fibres in the two types of cultures.
ACKNOWLEDGEMENTS
The author wishes to thank Sir Joseph Barcroft for the kind hospitality given to him in his Laboratory, and Mr E. N. Willmer for much advice, criticism and help in preparing the manuscript. The author’s thanks are particularly due to the Professional Committee of the Central British Fund for German Jewry for the allocation of a grant.