Wax Secretion in the Cockroach

Wigglesworth’s extensive work in 1947 and 1948 on the structure and formation of the cuticle has shown a sequence of processes by which an epicuticle is formed. In Rhodinius, it seems that during the deposition of the cuticle, a wax appears over the surface of a semi-liquid protein layer rich in polyphenols which will subsequently become the ‘polyphenol layer’ of the epicuticle. At this stage the ‘wax’ can readily be removed, but later it has become a harder material. In the waterproofing of eggs by the tick (Lees & Beament, 1948) it is clear that the wax, rendered water-dispersable in the reservoir of Géné’s organ by being linked with protein, is a mobile material completely devoid of protein by the time it is produced at the surface of the organ. This material will spread over the surface of a freshly produced egg, and for a short time afterwards it is sufficiently mobile to pass from one newly laid egg to another which has not been in contact with the organ. But it soon becomes a hard, non-spreading material, and it is important to note that when the organ is made to secrete, but with no eggs to receive its secretion, wax accumulates on its surface with much inferior spreading properties.

Even more remarkable in this respect is the plastic waxy substance with which the red spider mite Metatetranychus coats its egg (Beament, 1951). This can be shaped into the long spike which surmounts the egg, by the soft ovipositor of the female, as it is withdrawn. Yet after an hour, the extremely brittle wax has a high meltingpoint, and it is not long before it melts at 170° C. These are outstanding examples of the type of change which is believed to go on in insect waxes immediately following their secretion.

The starting-point of this investigation was a chance observation by Prof. V. B. Wigglesworth (to whom I wish to express my thanks). Grease obtained by washing the external surface of many cockroaches (Periplaneta), was stored in a cupboard for several months. Now in its natural state on the animal, this grease is a vaseline-like substance which has no true melting-point ; it ceases to be a waterproofing material at about 33° C. But the stored sample had changed to a solid hard material, looking very like a white wax. It had a melting-point of the order of 56° C., it was readily soluble in chloroform, benzene and similar solvents, and its general properties were very similar to those of, for example, the cuticular waterproofing wax of Rhodnius (Beament, 1945). When this material was deposited from chloroform on to a membrane of lipoid free locust wing in a membrane holder, it produced reasonable waterproofing, with a transition temperature of the order of 50° C. Generally speaking, therefore, it appeared to be a wax composed of long-chain paraffinic components. Now paraffins, and their related alcohols and acids, are known to be chemically extremely inert materials. Nevertheless, it seemed very necessary to examine all the possible processes by which a change such as this could come about. It is possible that slow chemical change, such as oxidation or reduction, could have altered the materials present; polymerization of shorter components might have produced a wax of larger molecular weight, and therefore higher melting-point : or some enzymic, or even bacterial, process may have gone on.

Accordingly, small samples of grease were obtained by washing living late nymphal cockroaches with cold chloroform, and allowing the solvent to evaporate for 24 hr. Chloroform is readily detected by its smell, and this was used as a crude but probably quite sensitive test for the freedom of the sample from solvent. (It is later shown that these samples might have contained traces of chloroform, but it is most unlikely that this would in any way affect the validity of the results, for one does not expect chloroform to combine with these natural greases, or cause them to change.) These grease samples were then placed in small sealed glass tubes, in atmospheres of nitrogen, hydrogen, oxygen and argon; some were stored in darkness and others in sunlight and, in some tubes, the volume of gas was very small. They were left for periods of several weeks before being opened and examined.

In no case did any hard wax form in these tubes. To check that the process of change had not proceeded at least partially, it would be necessary to determine the melting-point of the samples, but for such a grease this is not possible. It is known (Beament, 1945) that the transition temperature of the grease sample would be a more reliable indication of changes, but again, there is appreciable wastage in making up standard membranes, the amounts of material to be tested were very small, and some micro-method was sought. Now it is known (Ramsay, 1935) that a droplet of water, placed on the cockroach, is rapidly waterproofed by a film of the grease spreading over it. These circumstances cannot be reproduced on clean glass cover-slips which had been coated with grease by being wiped over the surface of a living cockroach; a water droplet, added to the cover-slip, was apparently covered by grease, but very soon made contact with the underlying glass, whence it spread rapidly through the glass-grease interface to the edge of the cover-slip and there evaporated. One needs a surface with different wetting properties.

Now if the glass is replaced by polythene, or polytetrafluorethylene sheet, with which water has a contact angle of 90°, waterproofing can be obtained, and a drop so covered will remain for a very long time at room temperature. If these plastic sheets are then sealed on to the end of a wide glass tube, the other end of which is open, and then suspended vertically so that the sheet is in the surface of a water-bath, the film can be heated slowly, while at the same time preventing its being exposed to the almost saturated air above the bath. The temperature must be raised slowly, for such plastics are very poor conductors of heat, but at the transition temperature, as shown by samples removed directly from living animals, droplets of i mm.³ disappear in about a minute. This gives a rapid and rehable method for determining transition temperatures for very small samples of grease, and the material from the sealed tubes was so tested. The results are shown in Table 1 ; this also includes comparative values for a few samples which were large enough to be checked by standard membrane methods. However, determination by membrane holder takes a number of days to perform, and causes the grease to be exposed to higher temperatures for considerable periods ; it is possible, in fact, that these values are not as reliable as those obtained from the droplets.

It will be seen from the table that some small changes have taken place in the transition temperatures of the grease samples in the sealed tubes, but none of them have even approached the figure of 50° C. although they have been stored for a period during which complete transformation to hard wax would occur in a freely exposed sample. From our original hypothesis we can certainly eliminate the possibility that direct oxidation is responsible. It also seems most unlikely that the mere operation of sealing the grease in a tube would have prevented polymerization, where every precaution was taken to avoid heating the material; equally well, reduction processes would not have been eliminated, and as no antiseptic precautions were taken, it is hardly likely that bacterial action would have been prevented in this way.

The only evidence of any possible significance is that the samples with the highest transition temperatures come from tubes of greatest enclosed gas space. Is it possible that evaporation is taking place? A sample was heated on a glass slide to ioo° C. ; the grease was rapidly transformed into one with a higher transition temperature. It was soon discovered that mere prolonged heating would produce a material which would not spread over water at all at room temperature. Subsequently, a large sample of grease kept for several hours on a glass plate at 60° C., changed to a ‘wax’ with a melting-point between 50 and 55° C. If a sample was rapidly heated to a much higher temperature (bearing in mind the possibility of structural change or decomposition which this treatment might induce) a vapour could sometimes be detected which carried the distinctive odour of cockroaches, but which was not so pungent as to give the impression that it was composed entirely, or even mainly, of this aromatic material ; equally, it was not chloroform. In further tests, grease samples were stored at room temperature under high vacuum, and again small but significant increases in the transition temperature were noted; a wax with a melting-point of 50° C. could be obtained in this way much more rapidly than by exposure at the same temperature to the atmosphere.

As a result of these preliminary experiments, it seems that the transformation of grease into wax is not—or is not chiefly—due to a progressive chemical process. Since the changes can be speeded either by heat or by evacuation, there is every reason for supposing that some material is ‘lost’ in the hardening process. One must presume that the loss of a volatile component of the original grease is the principle cause for the change, and that its evaporation at room temperatures is very slow indeed. Despite the diversity of products secreted by insects, it would be very surprising if they could produce any of the typical wax solvents used in the laboratory, such as chloroform, ether, benzene, etc., which are either anaesthetic or highly toxic substances to them. All one can reasonably assert here, is that such a volatile solvent would have to be a liquid at room temperature, and must therefore be of a low order of molecular weight, especially as compared with wax molecules whose carbon chain may be 24-32 atoms long. The determination of molecular weights of solvent and wax is therefore a valuable approach.

It seems likely, from the evidence given above, that most of the volatile materials present would vaporize in a vacuum at temperatures above 60° C. A sample of grease, obtained from chloroform washings of living animals, was allowed to evaporate until no smell of chloroform could be detected. It was divided into two roughly equal parts, one of which was weighed and then heated under reduced pressure at 64° C., thus giving an indication of the weight of volatile material present in the sample. The other portion was weighed and injected into the Torricellian vacuum of a mercury barometer, jacketed with methyl alcohol vapour at 64° C. From the volume occupied by the volatile part of this sample, its vapour density could be calculated : its mean molecular weight is of course twice this figure.

The results showed considerable variation; values for molecular weight varying from 100 to 200 were obtained. It is, however, obvious that if the sample is stored for a long time, solvent and chloroform are both lost, whereas the determination of molecular weight on a fresh sample must include some chloroform vapour. But even if one cannot make a direct suggestion as to the nature of the solvent, these experiments do at least provide concrete evidence for the presence of a volatile component in the grease. Clearly a different method for determining molecular weights is needed.

Molecular weights of materials in solution can be obtained by their depression of the freezing-point of a suitable solvent. Ramsay (1949) has described a method for the determination of very small samples of aqueous materials, and his method was adapted for this purpose. Benzene was chosen as the solvent ; it has a very large cryoscopic constant as compared with water, and is a good wax solvent. In Ramsay’s method the sample is stored under liquid paraffin, and a small sample drawn into a fine silica tube, enclosed in the paraffin. For the present purpose, a liquid insoluble in benzene is needed, and pure glycerine was used. Since benzene is appreciably less dense than glycerine, the samples were stored by injecting them into the highest point of an inverted U-tube of fairly narrow bore filled with glycerine, so that the liquids remained in by capillarity. A number of large nymphal or adult cockroaches were rapidly washed with recrystallized benzene to obtain the grease solution; this was carefully decanted from the collecting dish, for small amounts of watery materials—probably excreta—are often also washed off. The volume was carefully measured and half of the material was evaporated to dryness and heated at 60° C. to drive off the benzene and the solvent; samples of the other portion were stored under glycerine in U-tubes. The dried portion was weighed, taken up in a small quantity of benzene, its volume measured and similarly stored under glycerine. Samples were then drawn from either tube, and their freezing-points determined by Ramsay’s method. The freezing-point of pure recrystallized benzene, after storage in glycerine, was also taken, which by subtraction gives the depression caused by the wax or grease in solution. It is appreciated that any inter-solubility between benzene and glycerine might have an effect on the results obtained, but the values for pure benzene after storage in glycerine are very close to those quoted by the manufacturers, using a standard salt solution for calibration of the Beckman thermometer. Certainly, the concentrations of waxes used cause a large depression in freezing-point, as compared with the possible variation in the ‘pure’ benzene through dissolved glycerine.

From the depression of freezing-point of the sample containing the hard, nonvolatile wax component, its mean molecular weight was obtained. The other sample gives the freezing-point of a solution containing a known molecular amount of wax, together with solvent molecules. From this, one can calculate the mean molecular weight of the solvent, for various proportions of solvent to wax. The results are shown in Table 2. On p. 522, the amounts of volatile component present are given, and it is now possible to indicate the range of molecular weight possible for the volatile material; it must lie between 140 and 190. Estimates are also given for the molecular weight of beeswax by this method, and it will be seen that they agree very well with the results published by Chibnall et al. (1934).

A relatively small molecule, such as this ‘solvent’ would appear to be, might easily be identified by infra-red spectral analysis of the grease, supposing it to be markedly different from the molecular configuration of the wax molecules. However, an examination of grease samples taken directly from the surface of freshly moulted animals did not indicate the existence of anything other than long-chain saturated paraffins, with CH₃ and OH groups; there was a possibility of an aromatic ring compound. Grease could not be obtained in sufficient quantity to make this approach of further practical value.

If the cuticular grease does contain some volatile materials however, these would be evaporating continuously—though probably in minute quantity—from the surface of living animals. Accordingly, a culture of several hundred cockroaches was established in a glass tank, kept as clean as possible, and fed on sugar, dry bread and water, thus attempting to reduce to a minimum the volatile materials from the food. The tank was sealed, except for one air inlet through which air was drawn via a filter; at the other end, air was removed through a polythene tube and passed through a trap immersed in solid carbon dioxide. A further filter was used to guard against any oil vapour diffusing back from the pump moving the air stream. A small volume of a watery fluid was collected each day for about 3 weeks, and kept in a sealed container; it smelled very strongly of cockroaches. The odour could be removed by adding an absorbent material such as a clean piece of porous pot or activated animal charcoal; by subsequent chloroform extraction of the absorbent, and evaporation in a micro-distillation apparatus, a small quantity of an acrid, tarry substance was obtained. This appeared to have no wax-solvent properties when a small fragment of beeswax was added ; it was very slowly volatile, and it seemed most unlikely that it could be the component responsible for the dilution of the wax to a grease.

The remaining liquid appeared to consist almost entirely of water; it was repeatedly extracted with chloroform in a separating funnel, and the residues were distilled to a small volume at room temperature in a micro-distillation apparatus under reduced pressure. This material was again subjected to infra-red spectral analysis; apart from chloroform, it showed only the presence of long paraffin chains, CH₃ and OH groups.

Chibnail et al. (1934) have shown that two insect hard waxes consist of entirely saturated long-chain paraffins, alcohols and smaller quantities of acids and esters. Beament’s comparative study (1945) of beeswax with other epicuticular waxes indicates that most of these could consist of a similar series of compounds, i.e. the even-numbered members of the aliphatic series, covering a wide range from perhaps C16 to C36.

The evidence presented above indicates the presence of solvent materials of low molecular weight, compared with the hard waxes, but chemically identical with them. Now it would be most unwise to assume that such a solvent would consist of one discrete chemical entity, or indeed of a very few components. It is, in fact, more than likely that the solvent would be made up of the same chemical species which are considered to compose the hard wax, but that the various series extend down into the range where they are volatile and liquid at temperatures in the biotic range. The pertinent properties, for even-numbered carbon-chain members of the alcohol, paraffin and acid series are given below; it seems likely that these are the only ones which insects would produce, and that plants produce the alternate members of the series (Chibnall et al. 1936):

In addition to these properties, one must of course, bear in mind the water solubility of the compounds, for unless their lipoid solubility is most pronounced, they will not in all probability be useful in a grease directly in contact with a water-saturated cuticle. It is immediately seen that if the acids are part of the solvent material, they would have to be of very short-chain length to be liquid at biotic temperatures. They are then very water-soluble, and it is shown below that the lower members of the acid series are not good wax solvents.

It is further interesting that long-chain acids do not make up more than a few per cent of beeswax, according to Chibnall’s analysis, and that COOH groups were not recognizable from infra-red analysis. Perhaps most important is the very strong and characteristic smell of the lower aliphatic acids—this is certainly not detectable when the grease is heated, and weighs most strongly against the presence of acids in the solvent.

Of the alcohols and paraffins, hexanol is appreciably soluble in water, but octyl alcohol has a much greater affinity for oily materials ; decyl alcohol is the highest member of the series which would be sufficiently liquid at room temperatures to ‘liquefy’ a wax. On the other hand, higher chain lengths of the paraffins are still liquid, and even tetradecane is just liquid in the biotic range of temperatures. Hexane, on the other hand, is extremely volatile, and octane would seem the lowest member of the series to be of possible interest.

If it is true that in the cockroach, the manufacture of paraffins and alcohols extends in each series down to chain lengths sufficiently short to be liquids, then the production of a ‘solvent’ is not at all difficult to understand. Whether the grease contains similar amounts of each of a series, or whether there are larger amounts of particular members is not clear. It does, however, seem that there is a disproportionate amount of short-chain materials in the grease—on the evidence of molecular weights, and it is possible that they are derived from intermediates in the process of building up the long-chain waxes.

Further information can be obtained from the rate of evaporation of a chloroform extract of grease. If this is plotted against time, and compared with recordings of the loss of chloroform from a saturated solution of beeswax of similar surface area and under identical conditions, one can allow for the evaporation of chloroform from the grease, and thus obtain an indication of the pattern of the rate of evaporation of the solvent. Grease samples varied quite considerably in the amounts of volatile material which they produced. It was quite obvious that evaporation of two kinds took place : a larger quantity evaporating less quickly and representing up to 40 % by weight of the grease, and a more volatile one which might be 15 % in a solventrich sample.

The further investigation of this problem is made difficult by the lack of microchemical methods for the identification of waxes and the minute amounts of fresh materials available, especially as they must always be contaminated with the substance used to extract them. A comparison of the properties of the natural grease with those of a synthetic grease compound from a readily available wax—beeswax—and paraffins and alcohols of suitable length as solvents will provide another approach. Such a comparison would be more justified if even the hard waxy part of the cockroach grease could be obtained in larger quantity ; although what little evidence we have suggests that Chibnall’s analyses for beeswax do apply to other epicuticular waxes, the peculiarities of the cockroach grease might make it a little dangerous to assume that this also consists only of compounds such as are found in beeswax.

If a cockroach comes into contact with any porous and absorbtive material, some of its mobile grease may be lost. The cultures of cockroaches in the Zoological Laboratory at Cambridge normally contain a number of inverted flowerpots of unglazed earthenware, in nests four to six high, to provide a shelter for young stages, and a rough surface for attachment of animals during moulting. Cockroaches appear to avoid a clean flowerpot, placed freshly into a tank; they are rarely to be seen on one as compared with those pots which have been in the culture tank for a few weeks. An attempt was made to extract the less volatile components of the grease from such flowerpots. A number of pots were first thoroughly washed in chloroform, to remove any fat-soluble materials in them, and left in the culture. They were withdrawn after 3 months, and placed in a large glass vessel, where they were repeatedly washed with distilled water, removing a quantity of debris and no doubt a certain amount of water-soluble materials. They were then allowed to dry and were repeatedly extracted with boiling chloroform. The material obtained on evaporation of the chloroform was a brownish very soft waxy substance. From a sample of this about half a gram of a white wax was obtained by recrystallization from boiling acetone. The material had a melting-point of 53-56° C., as compared with 56° C. for the wax left by evaporation of the natural solvent in air. The very fact that such a large quantity of wax can be readily obtained from flowerpots can only be explained by assuming that it is the non-volatile component of the copious grease on the animal. It seems most unlikely that much oil-soluble material would come from the excreta of the animals, or from the food which is placed in the tank. This material is therefore considered to be the cuticular wax in the experiments which follow, but as an additional check, a sample of beeswax, similarly prepared by recrystallization, has been included in these tests. It will be seen that there are no fundamental differences in behaviour between the two substances.

It was first decided to investigate the properties of solutions of waxes dissolved in various combinations of octane, decane, octyl and decyl alcohols.

When either beeswax, or the cockroach wax obtained as indicated above, are added in excess to chloroform, benzene, carbon disulphide or diethyl ether, saturated solutions are obtained, in which the liquid portion is not greatly different in viscosity from the pure solvent. Any excess wax remains undissolved, and if evaporation is allowed to take place, wax is correspondingly thrown out of solution. An accurate figure for the solubility of the wax in the solvent can be determined. But no such figure is possible when any of the four supposed solvent materials are used. Both insect waxes appear to be completely miscible with each of these solvents, either alone or with mixtures of alcohols and paraffins. There is no such thing as a ‘saturated solution’, and if the wax is present in excess, a gelatinous mass is formed. The process of solution is slow at room temperatures.

A long-chain alcohol or paraffin as a solvent has thus a unique property in relation to these insect waxes, which are turned into greases similar to that found on the surface of the cockroach. Furthermore, as evaporation occurs, a grease made up of beeswax and octyl alcohol, for example, merely becomes more viscous; if left at room temperatures the melting-point of the mass rises and approaches that of the original wax asymptotically. Only a material miscible in all proportions can behave in this way. It is even more striking that a sample of pure paraffin wax, m.p. 56° C. does not form a grease in octane/octyl alcohol mixtures, but behaves as beeswax in chloroform, etc.

So far as beeswax was concerned, there seemed to be little difference between the behaviour of the paraffin or of the alcohol, or of mixtures of the two, though probably due to its greater viscosity, the corresponding alcohol formed solutions rather less rapidly; there was no significant difference in solution rates at 50° C. The cockroach wax, on the other hand, was certainly more soluble in a mixture of alcohol and paraffin than in either alone. The rate of invasion of a solid wax block is perhaps a poor indication of the affinity of the materials concerned, but it is most important to note that in a corresponding series of experiments, the saturated acids are not efficient solvents as compared with the alcohols and paraffins, nor do they form ‘greases’ when the wax is present in excess.

Octane, and to a lesser extent, decane, are very volatile materials at room temperatures ; octyl alcohol is by comparison much less volatile. Their rates of evaporation from a synthetic grease may however be expected to be very much slower, since they would have great affinity for the molecules of wax. Standard samples of greases containing various solvents were spread on sheets of clean aluminium foil and allowed to evaporate under controlled conditions at 20⁰ C. The rate of evaporation of natural grease initially obtained from chloroform washings was also recorded under these conditions, but a correction was made for the amount of chloroform evaporating, using as controls solutions of beeswax in chloroform and of beeswax in chloroform and octane. From regular weighing of the foils, the following conclusions were reached.

The rate of evaporation of octane when present in about equal proportion to wax, i.e. the same proportion as the natural solvent at its maximum, is still rapid as compared with the loss from natural grease under similar conditions. The evaporation of decane is appreciably slower, while that of octyl and decyl alcohols is well within the range of the natural materials. However, there is already evidence for believing that in the natural grease there are two components, one of which evaporates more rapidly than the other. A solution of beeswax in octyl alcohol and decane, in which the proportions are 4:1:4 shows a similar pattern of evaporation, and the loss of solvents is only about twice as rapid as the loss from fresh samples of natural material.

If the original assumption—that the solvent consists of a series of compounds of the two species, paraffin and alcohol, running continuously into the range where they are solid waxes—is correct, then the addition of higher members, such as dodecane, dodecyl alcohol, to the eight and ten carbon solvents would certainly give a pattern of evaporation almost identical with the natural materials. Unfortunately pure dodecane and dodecyl alcohol were not available.

The grease of the cockroach is found on a cuticular surface saturated with water; it is also known (Ramsay, 1935 ; Beament, 1945) that it will spread rapidly over water. Since the hard wax part of the grease cannot do this, the effect of adding solvent on the spreading of our synthetic grease on water is obviously of importance. In view of the evidence (Lees & Beament, 1948) for the dispersion of wax in water in Géné’s organ of the tick, by its association with a protein, one must also discover whether these materials could act as dispersants, for the wax in Géné’s organ seems likely to contain solvent materials, but no waterdispersion takes place once the linkage to the protein has been broken.

All the materials investigated are lighter than water; solutions of the waxes, either in octane or in octyl alcohol alone, merely form a floating mass. In neither case does vigorous shaking produce an emulsion. But a solution of wax in both alcoholic and paraffinic components behaves remarkably when added to water in a test-tube.

Providing the glass is thoroughly clean, so that its surface is fully wetted by water, a thin film of wax spreads up the sides of the tube, often to a height of half an inch. From the mobility and properties of this film, it would seem that it is entirely supported by a water film underneath it, against the glass. An explanation of this phenomenon is not immediately obvious. Octyl alcohol, with its polar ending and high spreading pressure on water, might tend to form an orientated film with the wax. Octane, being non-polar would not be expected to show these properties. But one must realize that the typical wax molecule in beeswax for example, is two, and even three times as long as the octyl alcohol molecule. Where the alcohol is present in comparatively small proportion it is likely that almost the whole of it will be held in the water film, and one could suggest that such a situation would not be sufficient to disrupt the whole length of the residual attraction between the longer wax molecules. This disruption is, perhaps, the part played by the paraffin component of the solvent.

It is obvious from this evidence that a more detailed comparison of the properties of the synthetic materials, as compared with the natural grease, could be made with a surface balance. This might also enable direct measurements of molecule size to be made. The apparatus used for these experiments is believed to be of a novel design, and a detailed description of it is given in the Appendix to this paper; it does, however, record surface pressures for given areas of film spread on a water-trough exactly as in the conventional torsion wire models.

The various materials were added to the water-surface in solution in benzene. Compression/area curves for samples of natural grease, cockroach wax, various synthetic greases and of stearic acid are shown in Fig. 1. It will at once be noted that cockroach natural grease is capable of exerting a very high surface pressure, and that its compression/area curve is not smooth. At a pressure of 35 dynes/cm. a break occurs; this is usually interpreted, in a ‘mixture’, as showing that some element of a compound film has been ‘squeezed out’ of the monolayer. It is particularly important to note that at about this pressure, a film of decane/decyl alcohol collapses ; it would presumably be reasonable to infer that such a solvent would have been squeezed out of a complex film at that point. On the other hand, the materials which are left, and which are presumably the longer chain components, collapse at a very high pressure, above the collapse point of stearic acid, for example, but very close to the collapse point of the beeswax and cockroach wax films. Solvent may be squeezed out at above 35 dynes/cm. pressure and the film collapses above 80 dynes/cm. It will at once be obvious that these pressures, on the living cockroach, would be quite sufficient to promote spreading over the whole animal, regardless of whence the grease and solvent appear, and over a wet surfaced cuticle as well as over any small droplets of water on it; it is significant that, because these droplets are rapidly prevented from evaporating, they must be covered by a compressed monolayer.

Films of cockroach grease, under the best experimental conditions, remain stable on the trough for at least 48 hr., and the decrease in their area is of the order of 2 % per day. The same is true for synthetic grease films. If this change is due to evaporation, then rates of loss from a monolayer may well be lower than from a corresponding thick layer on a piece of glass or tinfoil, when the exposed area is taken into account. It seems possible that the evaporation of solvent may be retarded through the polar groups’ attachment to the water surface.

If a powerful emulsifier, such as the polyethylene glycol, C. 09993, is injected in minute amount into the water beneath the film, its collapse is immediate and dramatic, but the sprinkling of an elutriate of activated charcoal on to the surface (large particles may not be supported) does not seem to alter the film at all. This is surprising, but again it points to the extreme affinity between the orientated grease, with its solvent, and the underlying water (cf. Holdgate, 1955).

Provided one knows the amount of material added to a surface, spread as a monolayer, the area it occupies under zero compression, and its density, one can calculate the average molecular length and surface area occupied by each molecule on a water surface (Adam, 1941). With the material used in these experiments, the greater difficulty is to measure the amount of material added to the surface of the balance. This can however be approached, by carrying out a cryoscopic determination of a sample of grease in benzene, before adding a part of the same solution to the trough. The cryoscopic reading allows a calculation of the number of molecules in the sample, and thus the area and depth occupied by them when arranged in a monolayer. Results are shown in Table 3.

First, it will be seen that a very reasonable figure has been obtained by this method for beeswax, bearing in mind the results of Chibnall’s analysis. That the ‘hard’ portion of the cockroach wax is of shorter average length, is not surprising if we are right in assuming that the range of compounds in the grease may run continuously from hard wax to liquids at biotic temperatures. There would in fact be no sharp dividing line between ‘wax’ and ‘solvent’. The mean figure for the length of natural grease molecules, as compared with a hard wax sample, supports the idea that the solvent molecules are of the order of C12-C14 in length. It must, however, be remembered that while the alcohol portion of the solvent may always be occupying its full area in this film, the paraffin part, having no polar group, may well not be arranged in the monolayer, but lie between the upper parts of the longer molecules. This may make the values obtained above rather higher than is the true state of affairs. It has not, on the other hand, been possible to determine whether the distribution of chain lengths within the grease is uniform, or whether there are two ‘peaks’, one at short lengths, and the other at a much longer aliphatic chain size.

Wigglesworth (1945), Beament (1945), and later workers, all demonstrated that the vapour of, for example, chloroform makes insect cuticle and cuticle membranes more permeable to water, at room temperatures. Beament (1945) has further recorded that if extracted cuticular waxes are deposited on suitable membranes, from solutions in chloroform, the wax layer so obtained is not of the same order of impermeability to water as a corresponding natural cuticle, when the chloroform has evaporated away. Wigglesworth (1945) showed that once a cuticle had been heated beyond the melting-point of the wax, it is then irreversibly changed to a more permeable state, as shown by its transpiration at biotic temperatures. An artificially waxed membrane (from chloroform solution), on the other hand, becomes slightly less permeable at lower temperatures, after the wax has been melted, and this is especially so if the membrane is kept water saturated. The wax from the tick Omithodorus (Lees & Beament, 1948) cannot be made to waterproof a membrane by deposition at room temperatures from chloroform, although a relatively more impermeable layer can be formed in an atmosphere saturated with chloroform at high temperatures.

In view of this evidence, it is of particular interest to discover what effect the materials believed to be natural solvents in the cockroach would have on natural waterproofing layers of the cuticle, and on less efficient layers deposited on wax-free cuticles from chloroform. The information required is best obtained by using insect wings as membranes, for Beament (1954) has shown that water transport through normal cuticle may be appreciably complicated by the phenomenon of asymmetry. Preliminary experiments showed that cockroach wings, either before or after chloroform extraction were not asymmetrically permeable in a water gradient. Some of the experiments reported here were carried out in the type of membrane holder described by Beament (1954) ; the apparatus has the great advantage that the membrane can be reversed, or otherwise treated without being subject to any kind of mechanical change. However, this apparatus is for liquid/vapour gradients, and the investigation was extended here by using an apparatus to measure transport in vapour/vapour gradients; it is shown in Fig. 2.

The membrane is confined in a duralumin cell, between rings of neoprene of circular cross-section lying in V grooves. This has the great advantage that the membrane cannot be split by increasing the clamping pressure. Neoprene, as compared with rubber, is largely unaffected by the vapour of any of the solvents used. The cell can equally well be used to record liquid/vapour permeability. The membrane-containing cell is clamped between the ground necks of two spherical bodied flasks, having taps on the bodies at right angles to the necks. The clamping frame work of electron metal (for extreme lightness) also carries two adjustable case-hardened points on a balance beam. Adjustment of the height of these points relative to the beam will bring the centre of gravity of the apparatus to the point of maximum sensitivity, just below the pivots. This adjustment must be made after suitable quantities of liquid have been introduced into the two bulbs, by means of pipettes passed through the bore hole of the taps. After counterweighting the apparatus to a balance point, a steel rod projecting from the beam is rested on the hook of a sensitive torsion balance, so placed that its distance from the pivots is the same as that between the centres of the two spheres comprising the bodies of the flasks. Transfer of liquid through the membrane will thus behave as though moving from the centre of one sphere to the other, and will cause the same change in moment at the measuring point of the torsion balance. The apparatus is therefore null-point and direct reading.

In these experiments, sulphuric acid was used as the desiccating liquid on one side, and an equal weight of water gave a ‘saturated’ atmosphere on the other. It is important in introducing these not to allow them to splash the sides of the flasks, or to run on to the membrane and its cell. After a short period for pressure equilibration at the chosen temperature, the taps of the two sides are closed, and recordings of weight change are taken every 24 hr. The taps are opened momentarily every week to equilibrate any pressure differences which might build up between the two sides. Following the determination of the vapour/vapour permeability of a membrane, one of the various solvent materials was added in equal minute quantity to both flasks. In this way it was hoped that the weight changes recorded would not include the transport of any solvent, which would be present as ‘saturated vapour’ on both sides. After such recordings, either the membranes, or the humidity controlling solutions could be reversed, the membrane evacuated to remove solvent, and then recordings repeated. (See Table 4.)

So far as any topical fat solvents are concerned—chloroform, benzene, ether, etc.—their vapour makes a normal insect wing cuticle more permeable, whether this be a membrane of the cockroach or of a hard-waxed insect such as Rhodnius. After exposure to the treatment, even if the membranes are kept in a high vacuum for several days in the hope of removing all traces of the vapours, their permeability is still permanently increased, presumably due to a permanent disorientation of their innermost monolayer. But the vapours of octane, and of octyl alcohol do not produce any detectable change in permeability, either when experiments are conducted in their vapour, or when the membrane has been evacuated following exposure to either. There can be no doubt of the ability of these substances to penetrate into the lipoid layer of the cuticle, and even if they transform this into a ‘grease’ it is still apparently quite as impermeable as the original wax. The results obtained with cuticles where wax has been laid down from chloroform on a lipoid-free insect wing membrane are perhaps even more significant. These become more impermeable to water, and approach more closely the order of impermeability of their natural counterparts after they have been exposed to the vapour of the aliphatic solvents. That this is not due merely to the formation of a layer of grease above the solid wax deposit is again shown by the impermeability of such membranes after they have been subjected to high vacuum to remove as much aliphatic solvent as possible. It would be difficult to consider the efficiency of octane as compared with octyl alcohol in these experiments, but there is no doubt that a mixture of the two is most efficient.

Thus again, out of the range of solvents used here, it is only the vapour of a long-chain paraffin (or alcohol) that does not upset the impermeable structure of a natural cuticular wax layer. Further, such materials appear capable of producing orientation in longer wax molecules, and of reorientating a lipoid layer after the vapour of other solvents has caused misalignment. It must be remembered however (see discussion), that this evidence has been obtained with membranes, one side of which is exposed to saturated water vapour, and which therefore might contain a large amount of water.

The whole of the foregoing experimental investigation into the composition of the grease of the cockroach and its properties has been conducted by physico-chemical methods, and physico-chemical models. It was therefore somewhat disturbing to find that a sample of the grease is a very strong reducing agent. In the original hypothesis, one has assumed that the composition of the lipoid materials followed the same pattern as that of beeswax, and that they were entirely saturated compounds. Since the behaviour of beeswax (which is entirely saturated) dissolved in saturated solvents, is so close to that of the grease, and as there are no signs of the phenomena usually to be associated with the ‘drying oils’ in the grease, it may be suspected that the reducing materials are secondary products, and not components playing any important part in the formation or composition of the grease. The hard white cockroach wax, obtained after recrystallization from the flowerpots in the culture tanks has no reducing properties; an attempt to determine an iodine number for a sample of this wax left no doubt that it was a saturated substance. A few initial experiments, using unsaturated alcohols of the order of C₈ chain length, again indicated that these were by no means as efficient as solvents as their saturated counterparts. A sample of grease obtained by washing large nymphs or adults with chloroform would reduce ammoniacal silver nitrate, giving a brownish deposit. It would also blacken rapidly in solutions of osmic acid. Following either treatment on the grease, if chloroform was added, a grey ‘solution’—probably a dispersion—was formed which would pass through a fine-grade filter-paper. The dispersion could not readily be aggregated by centrifuging. A sample of grease which has been obtained freshly and changed to a hard material by heating in a vacuum—so that no oxidation could take place—still has strong reducing properties. This is of particular importance for it must surely eliminate any possibility that the solvent is also the unsaturated material. It was first thought that polyphenols and quinones might be responsible, for Pryor (1940) has shown that these are the main precursors of the tanning of the cuticle of the cockroach. Polyphenols are much more lipoid-soluble than water-soluble, and it would not be surprising if they dissolved in the grease from the underlying epicuticle substrate which is so rich in them. But no reaction could be obtained with ferric chloride, which is a most sensitive test.

It was therefore decided to try to fractionate and test for reducing materials at each stage. A large sample of grease was extracted with chloroform from flowerpots ; the odoriferous material was removed by adding charcoal and filtering. Apart from the difficulty of obtaining large quantities by washing living insects, the pot extraction ensures that one does not wash reducing materials out from underlying living cuticle. From the brown residue, obtained after the chloroform had been evaporated, the wax was taken up in boiling acetone, and the brown residues were concentrated by hot filtration. This material was soluble in warm ethyl alcohol, and showed very strong reducing properties. About 100 mg. were obtained. On cooling, and allowing the alcohol to evaporate, such a solution would form a delicate pale brown skin over the surface of a dish. It was insoluble in the cold in many lipoid solvents, such as chloroform, diethyl ether, benzene and carbon disulphide, and only slightly soluble when they were boiled, though cloudy dispersions could be formed.

The properties of this material were to some extent reminiscent of shellac, and a series of tests were carried out comparing the reducing residue with a sample of pure shellac from Laccifer lacca. The solubilities, forms of precipitate produced when water or acetone was added to an alcohol solution of similar concentration, were so completely similar as to leave little doubt that this material was a form of shellac. A micro determination of the iodine number gave a reading of 12, which compares with 10 for a sample from L. lacea (Parry, 1948) ; the material was strongly acid which is true of shellac. Finally, treatment with warm methyl alcohol and 3 % hydrochloric acid, followed by solution in ethyl alcohol to which diethyl ether is added, produced a fine white crystalline precipitate, with a melting-point of about 140° C. Parry (1935) outlines this test for a dihydroxycarboxylic acid in shellac which has a melting-point of 149° C.

Shellac is a material whose properties are far better known than is its composition, and such analyses as have appeared (Parry, 1935, 1948) indicate that it is a very complex substance. It is best regarded as a ‘molecule for molecule’ secretion of carbohydrate—laceóse—with wax. The term laceóse certainly includes a number of compounds, at least four of which have been identified, in the secretion of L. lacca. The amount of wax in the shellac of that animal is given as 3-5 %, but one is left in doubt as to whether this is wax, integral to the shellac secretion, or whether at least some part of it is an extraction of the waterproofing layer of L. lacca. The literature does not give any indication of the part of the epicuticle with which the shellac secretion is to be homologized, but it is here suggested that it is the cement layer of the epicuticle.

In the cockroach, the cement (Kramer & Wigglesworth, 1950) is a thin sheet, possibly discontinuous, and floating in the grease. It is certainly understandable that some of it would be lost to the surface of porous material over which the animal crawls ; certainly no other source for the shellac in extracts of the flowerpots can be suggested. The properties of the laccose undoubtedly enable one to explain many of the problems already encountered by workers on cuticle composition (see particularly Wigglesworth, 1947, 1948). In 1945, Wigglesworth and Beament first indicated that whereas boiling solvents would readily remove the waterproofing layer of wax from insects with hard waxes, cold solvents were not efficient ; this has always been attributed to the ‘protective action of the overlying cement layer’. The problem became more difficult to understand through a later, unpublished discovery that a solvent, at room temperature, but boiling under reduced pressure was almost as efficient in removing the wax from a cuticle. It has also been recorded (Wigglesworth, 1948) that whereas one can see a thin sheet of material lift from the surface of the cuticle when broken down by corrosive materials, one cannot find such a substance in the solution of extracted waxes, though Beament (1945) has drawn attention to the unidentified residues in the extracted waxes of many different species, and, in particular (Lees & Beament, 1948), to the very unusual platelets of material which separate from the much lower melting-point wax on cooling a chloroform extract of cast skins of Ornithodorus. These cannot be made to recombine with the wax for re-waterproofing an artificial membrane and they have poor solubility in all but boiling solvents.

Now many of these problems can be solved if one assumes that the cement layer is a shellac. Presuming that only a boiling efficient wax solvent can penetrate into the wax component of shellac, it breaks up the cement by mechanical action, thus allowing the main underlying wax layer to be dissolved. The laccose is dispersed, and as we have seen in the cockroach, it is so finely divided that it will readily pass through a filter-paper; it would therefore eventually settle in the extract. The reaction of the carbohydrate portion of lac with ammoniacal silver nitrate (the argentaffin test of Lison 1936) is so similar to the reaction of polyphenols as to suggest that it might be the cuticular tanning material, in the absence of a ferric chloride test. It is no wonder, therefore, that if cement is generally composed of shellac-like substances it has not been obtained as an intact lamina by using a wax solvent. The effect of shellac on artificially waxed membranes might yield valuable additional evidence here, but in the absence of micro-chemical tests for laccose, the development of this thesis must be an extremely difficult matter. Nevertheless, it is so typical of nature that what one insect—L. lacca—produces in quantity, others produce in minute amount, and certainly no other source of these materials in the cockroach can be envisaged.

The oenocytes in the blood of, for example, Rhodnius which has a hard, cuticular wax, are cyclicly active cells (Wigglesworth, 1933). They appear to be responsible for the secretion of the lipoprotein and wax of the cuticle at moulting, but they are quiescent at other times in the moulting cycle, unless wound-healing and repair processes are going on. But in the cockroach, Kramer & Wigglesworth (1950) have described how these cells differ in their behaviour as compared with the ‘typical’ insect by becoming permanently re-incorporated into the epidermis at an early stage of the animal’s nymphal life, and apparently remaining active continuously. Are we to conclude that in the cockroach the wax is produced only discontinuously, at the moment in ecdysis where the typical insect is secreting its wax, and that these cells then feed further solvents to the surface, to make up for evaporational loss throughout the instar? A grease can of course be made up from a small number of compounds of intermediate molecular size, which are just solid at room temperatures, or by the admixture of hard waxes (larger molecules) with liquid solvents (smaller molecules). The latter is certainly the case in the cockroach and the duration of the adult stage, and very probably, that of the nymph, is sufficient to lead to a large amount of evaporation of solvent, with a detectable change in transition point ; yet preliminary work does not indicate that this takes place, and the replacement of solvent seems most likely. The continuous production of wax and solvent would equally lead to an increase in transition temperature, and the accumulation of large amounts of wax on the surface of the animal.

Before we consider the implication of wax solvents on the general problem of wax secretion, it is perhaps worth considering briefly whether the presence of a fluid grease has any particular significance to the cockroach. Fluid greases—particularly the egg waterproofing material of ticks (Lees & Beament, 1948) could be considered as biologically essential ; for the egg this is obvious, and for the adult it may well assist in re-waterproofing the surface after the enormous distortion caused by engorgement. But as we have every reason for believing that the cockroach secretes its grease over its entire surface, just as other insects do, there is no reason why its wax should be more mobile initially. If the animal is subjected to temperatures above its transition point, one can attribute some importance to the fluid grease as a survival mechanism, for the grease becomes permeable and evaporation ensues, with consequent cooling (Edney, 1953). When lower temperatures prevail, the importance of the solvent in promoting reorientation has already been demonstrated. While the advantage of instantaneous repair from abrasion is obvious, such a phenomenon is not widely found, and the cockroach does not live normally in circumstances where it is liable to this sort of damage. Again, though Wiggles-worth & Beament (1950) have shown that the ootheca is not waterproofed, and it now seems likely that the spreading of grease over this structure from the female, augmented by her repeated rubbing of the limbs over it, provides its only waterproofing, surely this is an adventitious process, and not one which requires a fluid grease to be present at all stages in the life of both sexes.

The functional importance of solvent materials in promoting the organization of a tightly packed waterproofing monolayer cannot be over-stressed, for it is this very organization which we hold to be responsible for waterproofing in the hard-waxed cuticles. This takes place on a water-saturated membrane, and must be contrasted with the observation (Beament, 1945) that the grease takes about 6 hr. to spread over a dry insect cuticle before a waterproof layer is produced. It is probably true to say that the presence of a water-saturated surface is a most essential part of this process, and that the apparent anomaly of depositing wax in the presence of water must now be considered a necessity—in the cockroach. Holdgate (1955), discussing the stability of the contact angle of grease films on the cockroach, demonstrates that the water in the substrate is an essential part of this process. We can only make the briefest comment on the likelihood of a solvent mechanism being present in other insects. The rapidity with which their waxes become solid (Wigglesworth, 1947, 1948; Lees & Beament, 1948; Beament, 1951) must surely indicate that any solvent materials present during the process of secretion must be very much more volatile than those in cockroach grease. But their presence would do more than merely ensure a properly orientated monolayer; it has always been a source of speculation that the waxes separate into a layer distal from the polyphenolic materials during secretion. It seems very likely that the spreading pressure of such polar materials would appreciably assist in the separation process, for from such a mixture they would always come to occupy the outer, surface position. A solvent might also provide the most acceptable explanation for the way in which waxes can pass through the hardened cuticle during repair processes.

An investigation has been made into the properties of the grease layer which waterproofs the cuticle of the cockroach Periplaneta americana. This grease changes slowly into a hard wax if stored in air. The melting-point rises during this process by 20° C.

It is shown that the grease consists of a hard wax (m.p. 56° C.) dissolved in a ‘solvent’ which is liquid at room temperatures; there may be equal amounts of wax and of solvent. It is suggested that the series of paraffins and alcohols which probably compose the wax extend into the short-chain lengths, C₈-C₁₈ to provide the solvent.

A micro-freezing-point method for determining the molecular weight of wax samples is described (based on Ramsay’s method for aqueous solutions) ; the molecular weight of the solvent lies between 120 and 170, and of the hard wax—300-350. A method is also outlined for collecting large quantities of cockroach wax.

Of a large range of lipophilic liquids, which dissolve waxes, only octane, decane, octyl and decyl alcohols are miscible in all proportions with insect waxes, and form synthetic greases with beeswax and with the hard wax of the cockroach. The surfaceactive properties of the natural and synthetic greases have been compared on a surface balance of novel design which is described in the Appendix. The two materials are remarkably similar ; the surface balance has been used to confirm the size of the solvent molecules.

An apparatus is described which enables the continuous measurement of water transport through an insect membrane in a water vapour/vapour gradient, and in the presence of other vapours. As opposed to the effect of the most organic wax solvents, the vapour of an octane-octyl alcohol mixture does not alter the permeability of a natural insect cuticle to water ; exposure of an artificially waxed membrane to this vapour and water increases its impermeability. Evidence is provided for believing that octane-octyl alcohol mixtures will improve the waterproofness of most ‘imperfectly’ waxed membranes.

The natural grease of the cockroach is a strong reducing agent, and evidence is produced for believing that the lipoid part is entirely saturated, but that it contains a small portion of substances resembling shellac. It is suggested that this represents the ‘cement’ layer of the cockroach cuticle.

The importance of spreading agents with wax solubility in the formation of insect waterproof layers is discussed ; it is suggested that the presence of a water-saturated substrate is an essential part of this process.

I am particularly grateful to Prof. V. B. Wigglesworth, C.B.E., F.R.S., who provided the material with which this investigation was started, for his constant interest in the work. Mr G. D. Glynne-Jones kindly provided me with a quantity of pure beeswax. I have had valuable discussions with many members of the Zoological Department, Cambridge, and wish especially to acknowledge the advice of Dr R. H. J. Brown, Mr M. W. Holdgate and Drs M. G. M. Pryor and J. A. Ramsay.

Following the original work of Pockels (1891), Langmuir (1917), Adam & Jessup (1926) and Lyons & Rideal (1929) have been mainly responsible for the development of ‘surface balances In all their models, the principle adopted has been that of a very delicate float against which the monolayer film exerts its pressure. Measurement is carried out by a null-point method, the counterbalancing of the pressure on the float by the rotation of the torsion wire on which the float is suspended. Workers in this field are agreed that the principle is not without its disadvantages : the torsion fibre is temperature-sensitive ; the calibration of the float involves loading it in a different way from its normal measuring constraint; the robustness of the float system is inversely proportional to the sensitivity of the fibre ; the eventual measurement is subject to the interpretation of a geared tensioning device. Further, this is a measuring system which is very sensitive to unbalanced pressure with respect to the centre of suspension of the float, and any such unbalance is going to alter the overall tension in the measuring system. It seemed possible that many of these objections could be overcome by using a beam and knife-edge principle for measurement. The apparatus described below is shown in Fig. 3, and was used to determine the surface pressure information which is given on page 526.

A frame, constructed throughout in 1 × 1 × | in. brass angle, rigidly brazed at all joints, carries the trough of water on the surface of which measurements are to be made. Three levelling screws enable the trough edge to be made horizontal. The frame also carries the main pivot mechanism : an agate plate, carried in an aligning cage, and operated by a lever system spring-loaded against a fine pitch screw. The pivot can thus be adjusted to a given height; the beam is then freed to swing by dropping away the side supports, but it is essential (see below) that the actual vertical travel of the beam, in this process, be as small as possible. It is also an advantage for the beam to have a small swing. The beam centring and dropping devices must be placed well away from the knife edge (as opposed to typical balance design) in order that it shall be completely free of interference for small angles of swing. All parts are constructed of brass throughout, to minimize corrosion in the damp atmosphere over the trough. The beam carries the float against which the film exerts pressure. Now a knife edge can safely carry a load in excess of 200 g., so that the float could be robustly constructed of a casing of 0·001 in. brass shim, filled and coated with pure 65° C. m.p. paraffin wax, bonded on to the smooth metal with a layer of dried dilute Bostick. It is supported by an entirely triangulated framework of thin-walled | in. diameter stainless steel tubing, attached so as to allow vertical and horizontal alignment of the float with the line of the knife edge. The float-supporting structure meets the float entirely above the water line of the trough ; thus the float presents vertical waxed surfaces at all points of contact with the water surface. The best barriers between the ends of the float and the side of the trough were found to be the thinnest tissue paper impregnated with warm pure petroleum jelly.

The beam is notched at accurate intervals, and forces are applied to the film by adding riders of known weight to the beam. A simple calculation of moments about the knife edge gives the force per unit length of float for a known load on the beam. The compression barrier is moved by a double screw and chain drive and the beam carries a plain mirror, to obtain zero positions by lamp and scale.

After cleaning and sweeping the trough, the beam counterweights are adjusted to obtain a zero under free swinging conditions. The balance is clamped, barriers placed in position, the surface checked for contamination and the sample, in its volatile solvent, applied with an ‘Agla’ micrometer syringe. A suitable weight is added to the beam, the balance arm freed, and the compression barrier screwed along until the light spot returns to zero ; the area of film is then measured. The rider is moved to produce a greater moment and the operation repeated, thus giving the normal force/area curve. Restriction of the vertical movement of the beam is essential, since it is well known that dipping a vertical surface through a compressed monolayer will produce deposition of films on it, and thus remove an area of film from the surface at each operation.

A properly constructed beam balance on a knife and plate, suitably adjusted for centre of gravity, will respond to a displacing force of one dyne. This effect can easily be displayed by the lamp and scale. A dyne at the beam is equivalent, in this model, to 0-45 dyne over an 18 cm. float, i.e. 0-025 dyne/cm. This is, of course, well beyond the required accuracy of measurements of the kind indicated below; it is perhaps more important to note that 4 g. at the beam counterbalances a film of 100 dynes/cm., which can be measured accurately within 1:1000.