The Biology and Behaviour of Ptinus Tectus Boie. (Coleoptera, Ptinidae), A Pest of Stored Products: VI. Culture Conditions

Ptinus tectus is now a very common pest of ware-houses in this country. It is easy to breed in large numbers in the laboratory, and it is therefore a convenient species to use, among others, for biological assay of insecticides. In such tests, it is highly desirable to be able to get repeatable results ; this requires the test insects to be of uniform constitution and in a standard condition. Some relevant information about locomotory activity, which may affect mortality, has been provided in previous papers (Bentley, Gunn & Ewer, 1941 ; Gunn & Hopf, 1942; Gunn & Walshe, 1942; Bentley, 1944); considerable progress has now been made in producing pure lines by brother-sister mating; the present paper is concerned with the culture conditions required to give uniform animals. The work is not complete, but it has to proceed less rapidly for a time owing to the absence of the senior author and it is felt desirable to make the data generally available forthwith and without much discussion of the literature.

The animals used are the descendants of the stocks mentioned by Bentley et al. (1941); they have now been kept for several years at about 25° C. in mass cultures. For the past three years they have been bred in tongue jars about 9·5 cm. in diameter and 4·5 cm. deep, as described by Ewer & Ewer (1942). The cultures are now kept in a constant-humidity room at 70 % R.H. Each culture is initially populated by thirty adults which are allowed to lay for 7 days on 50 g. of food. These adults are given soaked cotton-wool to drink from; this is removed at the end of the week. Broadly speaking, about 150 adults emerge from such a culture. The pure lines have not been used in the experiments described here, except in one test of variability.

The food used is ‘Artox’ wholemeal flour with 5 % dried brewers’ yeast. This flour is a blend of no. 1 Manitoba and English Red Wheats, freed from weed seeds, ground by old-fashioned millstones, and has nothing added to it or taken from it. We are indebted to Messrs Appleyards Ltd., of Rotherham, for this information. The food is brought into equilibrium with air at 70 % R.H. before use.

The humidity control of the constant-temperature room is a modified copy of that in use at the Biological Field Station at Slough. The sensitive element is human hair; when the humidity falls, a fan blowing over a 10 1. water-bath is switched on and the water is heated electrically. The water level is kept constant by inflow from an adjacent tank with a float valve of the lavatory cistern type. The modification consists of a valve circuit which reduces sparking and the consequent fouling of the mercury (Fig. 1). In experiments, as distinct from cultures, the animals are kept in sealed chambers with a further humidity control by means of sulphuric acid and water mixtures; the reasons for this are given in § III (c).

Temperature is controlled by bi-metallic regulators with Sunvic vacuum switches, both in ovens and in constant temperature rooms. Pressure capsule regulation is too sensitive to changes of barometric pressure to be reliable for this purpose. For the experiments at 15° C. a special unit^* was constructed. This consists of a 100 1. capacity refrigerator linked by air supply and return pipes to a similar chamber without any refrigerating unit. When the chamber temperature rises, cold air is blown from the refrigerator to the chamber by means of a Goblin vacuum cleaner motor and fan ^†this make being chosen because it is exceptional in that the air current is not heated by being passed over the motor. A fan inside the chamber, operated by a motor outside it, automatically switches on when the cooling current is not required, so maintaining a uniform temperature. A low-wattage heater inside the chamber enables constant temperatures very near to and on either side of room temperature to be maintained. The advantage of a thermostat cooled by cold air over one employing a liquid circulation is that the oscillation of temperature inside glass vessels in the chamber is small because of the low thermal capacity of air.

The equilibrium humidity of the food for experiments is adjusted as described by Ewer & Ewer (1942). The weighed small quantities of food used in many of the experiments are made up as follows. A number of spoons were made by knocking a hemispherical depression in a strip of flat brass. From these a set was selected to give the required quantities. In using them, the flour is stirred up to a standard extent, the spoon is.dipped into the flour in a standard way and the heap in the spoon is wiped off flat with another strip of brass. The spoon is then emptied into the experimental tube. At first, each lot of flour was then weighed, and those lots which diverged too much were rejected. With practice in the use of the method, however, it became unnecessary to do more than weigh samples. By this method, it is possible to get 230 mg. of flour in each of many tubes fairly quickly with a range of less than ± 4 %.

The method of obtaining eggs is described below in the section on oviposition. For the experiments on larval growth, the eggs are freed from flour and put in small glass vessels to hatch. They are examined about the same time each day and all larvae removed, so that all the larvae are known to have hatched during the previous 24 hr.

The end of developmental life is reckoned to be the time at which the adult emerges from the cocoon and arrives on top of the flour. The quantities of flour used were seldom large enough to cause much delay between these two events. A sharper endpoint would have been the date of formation of the pupa or the date of the last moult ; but both of these events occur inside a cocoon and often cannot be seen without breaking open the cocoon. The emergence from the cocoon takes place about a week after the last moult at 25° C., and this duration is no doubt complicated by variations in behaviour; but the appearance of the adult on the surface occurs without interference and can always be seen. Consequently the durations of development given below are always from hatching to emergence from the cocoon.

As soon as each adult emerges, it is weighed on a torsion balance (Fabergé 1938). Its sex is determined by separating the wing covers, squeezing the abdomen slightly, and observing the genitalia which emerge. Hinton (1941) has figured the male genitalia, which consist of a median lobe with a some what longer pair of lateral lobes. The female has a telescopic tubular ovipositor with a pair of short setose processes at the end. After examination, the animals are destroyed.

The choice of a temperature at which to keep Ptinus cultures—and therefore at which to carry out these experiments—is important. It is desirable to use a temperature at which development is rapid, so as to accelerate results ; it is undesirable to use the temperature at which development is most rapid, for the following reason. Temperature-development data are commonly plotted in two ways—speed of development (reciprocal of duration) against temperature, and duration against temperature (e.g. Ewer & Ewer, 1942). The former way gives an approximately straight line, showing that the speed increases uniformly with rising temperature, up to the critical temperature; this line then bends over and begins to go down. The latter way gives a curve which descends to the ‘optimal’ temperature, where it is horizontal, and then rises for a short distance. The ‘optimal’ temperature; at which development is quickest, is slightly higher than the critical temperature. At the ‘optimal’ temperature, the accelerating effect of the last increment of temperature is just balanced by some detrimental effect of that increment, so that there is no resultant acceleration. The so-called optimum is thus à temperature at which some detrimental effect is starting to intrude. It would therefore be unwise to conduct the experiments at the ‘optimal’ temperature if one wished to draw general conclusions from them.

Ewer & Ewer (1942) concluded that the critical temperature for P. tectus lies between 23 and 25° C., while Howe (1943) has quoted 22–5° C. as the optimum. As Ewer & Ewer (1942) pointed out, some uncertainty entered into their results because of difficulties with temperature control, particularly due to enemy action. Accordingly, the durations of larval development and also of complete development from hatching to emergence from the cocoon were determined again at various temperatures between 15 and 25° C. It was found that with ample food (230–250 mg.) and at 70 % R.H., the average durations of larval development and of complete development were some days shorter than those given by Ewer & Ewer (1942) for 15° C., longer for 20° C., and a day or so shorter at 24–7° C. The temperature and speed of development curves plotted from these figures give a critical temperature very close to 24–7° C., i.e. near the upper limit of the range suggested by Ewer & Ewer (1942) see Table 1 and Fig. 2). Ewer & Ewer worked with animals from cultures bred at room temperature, while our stocks have been kept at 24·7° C. for some years. This may have caused alterations in the stocks which, together with changes in the methods in use here since 1941, might account for the divergences. Thus the temperature of 24·7° C. is not too high for this work, but is the highest which it is proper to use. It is certainly very close to the critical temperature. Perhaps the slightly smaller size reached at this temperature than at 15 and 20° C. (Fig. 3) reflects this fact. In any case, enough experiments have been repeated at 20° C. to show the general validity of the results.

An experiment was carried out to find out what effects slight variations in humidity are likely to have. At 66, 70 and 76 % R.H., the durations of development from hatching to emergence from the cocoon were 54·3, 53·6 and 52·2 days respectively. The differences are in the expected direction but are insignificant for most purposes. Accidental variations of one or two per cent in relative humidity are thus not likely to be a considerable cause of variability in the main experiments. Higher humidities cannot conveniently be used because of the growth-of moulds (Snow, Crichton & Wright, 1944).

These experiments do not pretend to discover the pure substances essential for full and rapid growth (cf. Fraenkel & Blewett, 1943). They are intended to find out what common materials should be used in cultures to produce the best results. The quantity of food required was found in separate experiments (§III (d)) and in the experiments on quality reported in this section an adequate quantity was always present.

Newly hatched larvae were put into separate tubes of 25 mm. diameter with 230–250 mg. of the food being tested. They were kept at 24·7° C. and 70 % R.H. in closed chambers which were opened and ventilated mechanically twice a week. The room in which this was done was kept at the same temperature and humidity, so no disturbance of these conditions occurred when the animals were examined and ventilated. It is advantageous to use closed chambers because this reduces the danger of infestation by mites. Such infestation leads to prolongation of development and to a raised mortality, so experiments with any infested tubes were discarded and ignored in the results given.

Two tests of suitability of the diet were used, namely the duration of development and the weight of the newly emerged adult. The duration of development included larval and pupal duration and also the period from the last moult to the emergence of the adult from the cocoon, which lasts just over a week. Fraenkel & Blewett (1943) did not count this pre-emergence period, and so an adjustment of about 8 days (Ewer & Ewer, 1942) is necessary in comparing their data with ours.

It will be seen from Table 1 that when the results are arranged in order of increasing duration of development, they fall roughly into the order of decreasing weight of adult. That is to say, if the criteria of good food are short development and large adult, then the two criteria lead to the same conclusion about the value of the diet. Wholemeal flour with 5 % dried yeast (now the standard diet) was used as a control for the other diets, and so appears several times in the table. This combination appears to be practically as good as a diet containing more dried yeast and better than one containing more protein and fat (fishmeal). The weight of adults from fishmeal is rather higher than would be expected from their duration of development. Evidence was obtained from experiments with inadequate quantities of food that a mixture of 50 % Artox: yeast is probably and 100 % dried yeast is certainly less satisfactory than 5–20 % dried yeast and the rest Artox. With 20 animals in an experiment, it is not unusual to get variations of average upto ± days and ±0·15 mg. Consequently variations of this magnitude on the various foods are not necessarily of importance. This is consistent with the results of statistical analysis.

The lower part of the table-shows the result of an experiment using a sample of a proprietary brand of cocoa and dried milk which was sent in infested with P. tectus. It is at once apparent that this material is not a good diet, even when supplemented by yeast. Nevertheless, the tin contained 331 dead adults, 72 live ones, about 100 pupae and 507 larvae. In the table the experiments marked ‘cocoa stock’were done with the progeny of animals from the tin. It will be seen that on standard food they developed as quickly as animals from the laboratory stock, while on the cocoa mixture they developed more slowly. The differences between the stocks were not statistically significant. There is thus no support here for Reynolds’s (1943) conclusions on the effect of parental feeding on the duration of development of TriBolium destructor.

One experiment was done to find the effect of a supplement of a salt mixture. The addition of 5% of Osborne-Mendel salt (Osborne & Mendel,1913) to 90 % wholemeal flour and 5 % yeast had no beneficial effect on duration. On the other hand, the same supplement to wholemeal flour without yeast was certainly beneficial, and this suggests that the dried yeast supplies whatever salt is necessary and absent from wholemeal flour. On wholemeal flour, perhaps some of the value of the yeast lies in the salt it supplies, rather than entirely on its vitamin content.

In these experiments, the standard food (95 % wholemeal flour + 5 % dried brewers ‘yeast) was used throughout and experiments were carried out at 15, 19 ·7 and 24 ·7 ° C. The results are shown in Fig. 2. It will be seen that the duration of development was not adversely affected until the quantity of food was very small. Indeed at 7 mg. of food there was a slight acceleration of development at all three temperatures, but with smaller quantities (3 mg.) there was a considerable retardation. Mortality did not usually rise above the common values of 0 –10 % until the amount of food was as low as 4 mg. In 2 ·8 mg. of food at 20 and 25 ° C., the mortalities were 20 %. At 15 ° C. one larva died, 11 emerged,-and 8 were still larvae 6 weeks after the last emergence; they were discarded. No animals completed development when allowed only 1 ·3 mg. of food.

An average Ptinus tectus brought up in good conditions weighs over 3 mg., yet animals can be reared without high mortality on under 3 mg. of food. They then sometimes weigh as little as 0 ·8 mg., averaging 1 ·1 mg. (Fig. 3), and they do not live long as adults. They must make highly efficient use of the food in body building. In fact, on small quantities of food the animals eat it all quite quickly and then feed on the tangled threads of faéces. It is not known whether this has any functional similarity with refection in the rabbit; certainly the faeces contain a high proportion of carbohydrate, even more than the original food.

The weight of the adult is affected by insufficiency of food when there is a comparatively large amount available for the larva. Fig. 3 shows the average weight plotted against the amount of food on a logarithmic scale. The improvement in size due to increasing the quantity of food is not great above 30 mg. per animal, while 60 mg. of food produces as large an animal as a much greater quantity.

It will be seen that as long as the quantity of food was 40 mg. or more, the lowest average weight reached was higher at 15 and 19 ·7 ° C. than the highest average at 24 ·7 ° C., but there was no consistent difference between the weights at the two lower temperatures. It is not uncommon for animals growing at lower temperatures to be larger than the same or related species at higher temperatures. For P. tectus any such relation seems to be effective only near the critical temperature.

Below 25 mg. of food per animal, temperature has no clear effect on weight of adult. Plotted logarithmically, all the points fall reasonably well on a straight line which passes near the origin. Such a relation is commonly found with drugs etc. when equal increments of dose produce diminishing increments of effect. Thus while 4 mg. of food produces animals averaging 1 ·2 mg., in six times as much food the animals weigh less than three times as much.

Some interest attaches to these results from the point of view of the efficiency of body building with limited amounts of food. Generally speaking, newly emerged females are heavier and have longer elytra than males. With ample food at 70 % R.H. and 24 ·7 ° C females average 3 ·42 mg. and males 3–12 mg. With an inadequate amount of food the differences are disproportionately less. Fig. 4 shows them averaged for the three temperatures used. It is at once clear that with quantities of food which are entirely eaten up, the females always have a greater average weight than the males, and that they therefore somehow use the food more economically in body building. This may mean that they use less of the food for basal metabolism or for locomotion, for there is no difference between the sexes in duration of development. It is noteworthy, however, that Park (1936a) found that adult females of Tribolium confusum Duval are larger than adult males and that they consume oxygen slightly faster than males when motionless, even when the rate of metabolism is calculated per unit weight. It does not follow, of course, that larval female Tribolium respire faster than males. In the case of Ptinus we are concerned with larval metabolism.

The difference of size is greater with larger quantities of food. This may mean that the female eats more when more food is available or it may mean that the male is still less economical—e.g. is still more active—when it can get more food.

As already suggested, the smaller average size of both sexes at 24 ·7 ° C. may be due to a detrimental effect of the temperature—if large size is the only criterion, it clearly is detrimental. Here again, the effect may be an indirect one, acting through its influence on behaviour (eating and locomotion). Since larger size is associated with lower temperature only when there is plenty of food, it is likely that larger size is partly due to eating more.

Estimates of the quantity of food eaten by an insect during development have seldom been made. Fraenkel & Blewett (1944) provided food of known water content and dry weight to newly hatched larvae of some stored products insects and then found the water content and dry weight of the residue after pupation. This showed the ‘amounts of dry matter and of water which had disappeared, but of course some of the dry matter at the end was faeces. In Ptinus the larva eats its own faeces, especially if short of food, so the amount of food eaten could be found only if the faeces could easily be separated from uneaten food and if food was replaced so frequently and abundantly as to reduce the consumption of faeces to a negligible amount. In our experiments the total amount of food passed into the alimentary canal has not been found; what has been found is the minimal amount of food which must be initially provided to produce an adult of maximal size. We do not know how much of this is eaten more than once.

So far, in § § III (a)-(d), we have considered those components of the physical environment which can be fixed in advance. Temperature and humidity remain unaltered throughout the experiment, provided the flour is not deep enough to allow gradients to be set up ; but the quantity of food gets steadily less as the animal eats it and the quality alters too, possibly because of selective feeding and certainly because of deposition of faeces. Contamination by faeces might be somewhat relieved by having the bulk of the food small, so that gaseous materials could escape freely, though of course the concentration of solid faeces would be larger. It would therefore be of interest to find the effect of shape of culture—i.e. varying depth with constant weight— on development; but the quantity of food required for a single animal is so small that the standard quantity must be put into very narrow tubes to get any considerable depth, tubes so narrow as to hinder the movements of the animal..Experiments were therefore carried out with several animals together, instead of one animal in a tube. The results suggested, that as soon as two or more larvae were put together, some quite new influence was intruding. Accordingly experiments were designed to see whether a group of two animals together in 460 mg. of food behaved differently from two animals each separately in 230 mg. of food. It was soon clear that animals in groups affect each other considerably.

Fig. 5 shows the average weights and average durations of development of groups of Ptinus varying in size from 1 to 32. In one set of experiments carried out at both 20 and 25 ° C., 1280 mg. of standard food were put in each of a number of tubes 2 ·5 cm. in diameter, with newly hatched larvae varying from 1 to 32 per tube. Thus the quantity of food per animal varied from 1280 to 40 mg. In another set there were 230 mg. of standard food per animal and in another 40 mg. per animal. In another set, tubes of 7 mm. diameter were used with no mg. of food per animal. As the graph shows, in all these experiments, when there were several or many animals together in one tube, the duration of development was prolonged and the weight of the adult reduced.

Clearly these effects of breeding the animals in groups instead of singly are not due to a high concentration of animals per gram of food, for they show clearly with 8 animals having 160 mg. of food each (cf § III (d)). They are not due to large bulk of food, for they are conspicuous with 4 animals in a total of 160 mg. of food. They are not due to small surface area of food relative to bulk, for one animal per 7 mm. tube with no mg. of food per tube produces substantially normal animals, while 4 animals per 25 mm. tube develop slowly and become undersized, although the surface, area is 13 times greater and the depth of food is in one case the same (1280 mg. of food) and in the other case only one-ninth (40 mg./animal) of that in the 7 mm. tubes.

A striking feature of the group effect is that it reaches its highest expression on duration when there are no more than 8 animals in a group. Thus at 24 ·7 ° C. the size and duration of development of animals bred singly are 3 ° 3 mg. and 53 days. With animals bred in groups of 8 –32, the figures vary somewhat around 2 ·8 mg. and 71 days, while for animals from groups of 3 –5 the values are intermediate. For 20 ° C., the corresponding figures are 3 ·5 mg. and 73 days for singles and 2 ·7 mg. and 109 days for groups. The full group effect is thus an increase in duration of development of say 40 % and a decrease in weight of about 20 %.

The group effect is also found in larger open cultures and not only in the tubes in closed dishes as described above. Thus in tongue jar cultures open to the air in the room at 25 ° C. and 70 % R.H., with ample food the average durations of development and the adult weights reached were much the same as in groups in tubes (Fig. 6). It looks therefore, as if the group effect is in some way due to the direct effect of one animal on another. It can hardly be due to ‘conditioning’ (i.e. contamination) of food, as described for Tribolium by Park (1938), for it shows clearly when the quantity of food per larva is constant.

The mortality in group experiments is much higher than with animals brought up singly, but there is no trend with increasing size of group between 8 and 32, such as Park (1938) found in somewhat different experiments. The figures for single larvae are 6 % ; for groups of 3 –5, 19 % ; for groups of 8 –32, 33 % ; and for tongue jar cultures of 20 –200 larvae with 230 mg. of food each mortality was 25 %.

An examination of the data was carried out to find if breeding larvae in groups affects the variability of weight and of duration of development. It was found that the individual durations of development of animals from groups were just as closely distributed about their average as were those of animals bred singly about their average. On the other hand, the variation in weight was much greater for groups than for singles. Fig. 6 shows that the largest animals are almost as big from groups of 8 –32 as from single tubes. The size of the smallest animals, however, is much smaller for groups than for singles.

Fig. 6 suggests that in groups there is some component of the environment which reveals the inherent variability of the animals, a component which is absent from the environment of an isolated animal. This hypothesis will be tested using stocks inbred by brother-sister matings; such animals inbred for 12 generations are not less variable than mixed stocks if the comparison is made entirely between individuals brought up singly (Fig. 6).

The group effect is apparent when there is quite a large amount of food, e.g. with only four animals in 1280 mg., and with 10 or 30 animals with 230 mg. of food each. Presumably, however, there must be a concentration of animals so low that the group effect disappears. Accordingly experiments have been carried out with very low concentrations of animals per gram of food and also with very large surface areas of food per animal.

The surface area of the food was varied in one experiment by using 8 g. of standard food and in another two experiments 0 ·5 g. in containers ranging in diameter from 0 ·7 to 30 cm. With 8 animals per container, no consistent effect of the consequent variation of area and depth was found. With the largest dishes the animals developed slightly more quickly but they were not heavier; this may have been due to partial virtual isolation of the individual larvae, for the food could not be spread quite evenly and lay in patches.

Varying the quantity of food available did affect the result. Fig. 7 shows the results of all experiments with groups of 8 animals at 25 ° C. with more than i mm. depth of food. It will be seen that the group effect does disappear if enough food is provided, but the quantity required is not less than 5 g. per larva, a hundred times the requirement of an isolated larva. This makes it quite clear that the group effect is in no way due to lack of food, which in any case does not prolong development unless the shortage is severe (4 mg./larva).

It is difficult to see how ‘conditioning’ or contamination of the food by the animals themselves could have such a striking effect, especially because an isolated animal in 40 mg. of food develops to normal size in normal time. Nevertheless, an experiment was done in which the food was changed weekly, with suitable controls. No significant effect on weight or duration was found.

An experiment was done to see how close the association of larvae needed to be to produce prolongation of development. Partitions of phosphor bronze gauze, of mesh too fine to permit the passage of newly hatched larvae, were put into 2 ·5 cm. diameter tubes so as to divide them equally into four. With 230 mg. of flour per animal, 4 animals per tube separated in this way emerged about 4 days sooner than animals not separated but otherwise given identical conditions. Certainly the group effect did not disappear completely or even largely.

The following observations extend those of Ewer & Ewer (1942) and are of some practical importance in the design of culture methods. In all cases, the insects were kept on National Straight Run English biscuit-making flour (N.S.R.) with casein and yeast, with particles small enough to go through a wire sieve having 100 meshes to the inch (say 40 per cm.). No advantage is claimed for this mixture as food over wholemeal flour + 5 % dried yeast, but the former is easier to reduce to suitable fineness all experiments have been done at 24 ·7 ° C. and 70% R.H. The eggs were collected by gently brushing the flour through a 60-mesh sieve, which retains them. During the sieving, the animals were normally allowed to drink water. It is very important to prevent flour and water coming together on to the animal ‘s body, because a hard cake is formed which frequently blocks the mouth and the posterior openings. Accordingly the animals are freed from flour by shaking in a coarse sieve before dropping them on to sodden cotton-wool. Movements of the mouth parts suggest that they then drink and they do in fact gain in weight. When they have walked off the wet cotton-wool, they are transferred to dry cotton-wool, to remove any droplets on the mouth parts or body surface, and only then are they put back on to flour.

It was stated by Ewer & Ewer (1942) that the female becomes mature between 3 and 12 days after emergence from the cocoon, though insemination may take place within 24 hr. of emergence. Those facts were determined by dissection. The age at which fertile eggs are first laid has now been determined. Females which had emerged during the previous day were mated either with newly emerged males (6 pairs) or with males which had emerged 10 –15 days before (7 pairs). They were allowed to drink each day. The two batches gave similar results (Fig. 8 A).

It will be seen that a few fertile eggs were laid within 3 days of mating, i.e. in less than 4 days after emergence from the cocoon. The full rate of laying (5 ·5 eggs/female/day) was reached on the 7th (5 ·2) or 8th (5 ·5) day after mating. The eggs were fertile from the start.

An experiment was also done with 4 pairs in which the males were newly emerged and the females 10 –12 days old (Fig. 8 B). These virgin females, which were allowed to drink each day, had been laying 2 –4 eggs per day, very irregularly; none of these eggs hatched. After mating, two of them laid.at an average rate of only 3 eggs per day, few of which hatched at first; eggs laid after the 17 th day showed a good hatching rate. The other two females laid 8 eggs between them in the first 24 hr., none of which hatched, 11 in the second day (2 hatched) and thereafter at an average rate of 5 ·0 per day, with a normal hatching rate (about 85 %). It is surprising that the fertility of the eggs was low at first, for remated females lay fertile eggs on the first day (see below) while insemination can be performed within 24 hr. of emergence (Ewer & Ewer, 1942).

The rate of laying recorded here and also by Ewer & Ewer (1942) is, of course, far higher than the 10 eggs in a lifetime mentioned by Fahmy (1931). In the first 3 weeks after emergence one female laid 11 eggs on 1 day (with a total of 41 in the surrounding week) while two others laid 10 each on 1 day (and 52in the week). Such high rates are only obtained if the animals are allowed to drink water daily. This is not surprising for two reasons: first, Ptinus tectus does not eat well unless allowed to drink (Ewer & Ewer, 1942) ; second, an egg obviously contains a high percentage of water and weighs about 0 ·034 mg-(1 % of the weight of the adult). Consequently the deposition of 5 eggs per day must be a serious drain on the water resources of the female. Rough figures indicate that when allowed to drink regularly, A female will lay 5 –10 eggs per drink.

When allowed to drink daily after weekly drinking, an average of 1 ·6 eggs were laid on the first day, 2 ·4 on the second and 5 or 6 daily thereafter. The dependence of a good rate of oviposition on availability of water to drink is now well established.

The influence of the presence of a male on egg laying is considerable. Five females were separated from their mates. They had been laying 5 ·4 eggs per female per day, of which 83 % hatched. After separation, the rate dropped to 2 ·8 (73 % hatched) in the first week, 1 ·1 (51 % hatched) in the second and 1 · 1(3 %=one egg hatched) in the third. Upon remating two of the females, 7 eggs were laid in the first day and 5 of these hatched, while the average for the first 5 days was 4 ·0 eggs per female per day of which 82 % hatched. This experiment shows that fertile eggs can be laid less than 24 hr. after the insemination of the females and leaves open the question of why mature females do not lay fertile eggs for more than 24 hr. after mating with newly emerged males. The slow start made by the two females mentioned above, which were not mated until 10 –12 days after emergence, may be simply an extreme variant or may be due to some deficiency of behaviour.

Seven of the pairs used in these experiments were kept until one member of the pair died. In the later part of the period they were allowed to drink each third day, when the eggs were collected, except when that day fell on a Sunday ; the interval without water was then 2 days. Under these conditions, in 83 days they laid 1124 eggs, a rate of 1 ·94 eggs per female per day. During this period and during the earlier period under various experimental conditions, each pair produced an average total of 647 eggs (501 –803), both members of the pair survived for an average of 276 days (213 –370). One female was remated after the death of her mate. She lived 376 days as an adult and laid 960 eggs.

Experiments were done with batches of 30 animals, which were given water to drink on 6 days per week. Five such batches laid an average of 3 eggs per animal per day when the food was fresh each day, but the rate fell to 1 ·6 eggs per animal per day on food which had been used for the same purpose several times before, and therefore contained faeces (cf. Park, 1934, 1936 b). The sex ratio was not determined exactly because some of the animals died a natural death and the sex is then difficult to discover. The viscera dry up some days before death and the genitalia cannot easily be squeezed out for inspection. There were 62 females and 37 animals of undetermined sex in the total of 150. It is evident, however, that with 30 animals on an area of some 65 sq. cm., the rate of egg laying per female is not far below that for isolated couples. The extreme possible values are 4 ·4 and 7 ·0 eggs per female per day and the probable value 5 ·6 eggs per female per day. At this density, crowding seems to have little effect on fecundity (cf. Park, 1933).

The usual method of breeding Ptinus tectus is to put 90 g. of wholemeal flour into a 2 lb. (approx, 1 kg.) jam pot, with a water supply for the adults consisting of a glass tube, 5 cm. long and 2 ·5 cm. in diameter, three-quarters full of water, with a piece of filter paper projecting from the corked tube. The culture is started by 30 adults 1 –3 weeks old, and they are left on for 7 weeks at 25 ° C. Certain other conditions need not concern us.

We investigated four jam-pot cultures of this type, in which 5 % of dried yeast was mixed with the flour; 739 adults were obtained spread over 6 weeks and individually weighed on the day. of emergence. When adults had almost stopped appearing the flour was sieved and 209 larvae were recovered. The distribution of weights of adults is dealt with above (Fig. 6). The average duration of development cannot, of course, be estimated with any accuracy.

By this jam-pot technique, 360 g. of flour produced 739 adults and could have, if more time had been allowed, produced 900 –1000. Even allowing for the fact that one culture became heavily infested with Psocids, so that only 123 animals emerged, this means that each animal had about 3301 –440 mg. of flour, or about ten times the minimum required by animals brought up singly. It has been shown above, however, that the group effect occurs unless a very much larger amount of flour than this is allowed each animal.

Thirty adults are capable of producing at least 3000 fertile eggs in 7 weeks, in good conditions, compared with the average of about 200 adults per culture which actually emerged. There is clearly room for improvement here. No doubt a large part of the deficiency was due to a fall in the rate of egg laying because of the conditioning of the flour by the adults, and there would seem to be no advantage in leaving the parental adults in the culture for so long. Indeed, a decline in the rate of egg laying to half its normal value is shown by animals replaced on used food for a second day or given food used by other animals for 1 day. It therefore appears best to allow a large number of animals in good condition to lay eggs in the intended culture for only 24 hr. They can be brought into good condition by a daily drink and a daily change of flour for 7 –10 days, provided they have not been long on contaminated flour.

In the routine tongue jar cultures, in which 30 animals are left on 50 g. of flour, with water to drink, for 7 days, about 150 adults are commonly obtained. By this technique, it is to be expected that the adults will emerge over a much shorter period than they will when adults are left on for 7 weeks.

Some experimental tongue jar cultures were run, in which newly hatched larvae were added to whole meal flour plus 5 % yeast at the rate of one larva per 230 mg. of flour. There were 6 cultures, two with twenty animals, two with 60 and two with 200. The results did not differ much, so they have been combined. The mortality was 25 %, the duration of development about 70 days, as in the group experiments, and the distribution of weights is shown in Fig. 6 

One test has been carried out with a line inbred by brother-sister mating for 12 consecutive generations. The newly hatched larvae were reared singly in 2 ·5 cm. diameter tubes, under standard conditions. Their average duration of development was 51 days; it is evident that this inbreeding has led to a reduction in size (Fig. 6).

Fig. 6 is designed to show the variation in weight of newly emerged adults under these various conditions. The graph shows the numbers of animals concerned and the standard deviation. Considering first the animals reared singly, it is at once clear that the inbred animals are small and that, under these conditions, they vary in size just as much as animals not inbred. Had they been brought up in a mass culture a reduction in inherent variability due to inbreeding might have become apparent.

On the other hand, the jam-pot cultures produced animals of large average size but of considerable variety of size. The mean temperature may have been slightly below that intended, and this could account for an increased size. It is not worth while to discuss the causes of the variation until analysis has gone further. The varied size in the group experiments cannot be explained by the varied conditions in amounts of food, etc. (cf. Fig. 5) ; there is a great variation with standard amounts of food the tongue jars gave a considerably smaller variation in size than the jam pots. This suggests that there is no need for the deeper container.