1. Under the conditions of the experiments described, Locusta migratoria migratorioides shows a preference for dry air in all parts of the humidity range, although dry air is by no means optimal for development, maturation, and breeding.

  2. The strength of the reaction is correlated with the magnitude of the humidity difference available, but appears to be little dependent on the region of the humidity range.

  3. The mechanism of the reaction is hygrokinetic and possibly hygrophobotactic as well, but probably not hygrotropotactic.

There is some field evidence that the distribution of locusts is related to humidity as well as to temperature. Lean (1931) has suggested that air humidity limits the area of dispersal of Locusta swarms, and similar conclusions have been drawn by Smee (1936) from the behaviour of the Red Locust (Nomadacris septemfasciata). Bodenheimer (on Schistocerca, 1930) and Hamilton (on Locusta, Schistocerca, and Nomadacris, 1936) in an extensive series of laboratory experiments, have shown that air humidity is an important factor in development and sexual maturation. Of the few laboratory experiments on the reactions of animals to air humidity only those of Rubtzov (1935) and Key (1936), discussed below, have been done on locusts. An investigation has therefore been carried out of the reactions of Locusta to air humidity as a stimulus.

The locusts used for these experiments were all of the gregarious phase. Stocks were obtained from time to time from the Locust Laboratory of the Imperial Institute of Entomology, British Museum (Natural History), London, and had originally been sent to the Institute from the Sudan and Kenya. They were bred in this laboratory in cages similar to those described by Hamilton (1936), and in larger ones, under the conditions of temperature and humidity prescribed by that author. It must be emphasized that the values given below for cage humidity have only relative meaning, since in all cases the air was almost saturated near the grass supplied as food, while over the heaters it was very dry. The hygrometer was placed approximately midway between the grass and the heater, with the sensitive paper of the hygrometer an inch or two from the floor. Humidity control was effected by varying the amount of moisture in the sand on the floor of the cage, and by varying the ventilation.

Under standard conditions

The alternative chamber, as described by Gunn & Kennedy (1936), was used for finding out whether or not locusts have a humidity preference. The animals could walk in this chamber, but the roof was too low to allow either flying or hopping. Under illumination of 65 metre-candles the locusts were too active for position readings to be made. In order to reduce the activity a sheet of brown paper was placed on the lid which kept the intensity of illumination at 5 metre-candles, and reduced the locusts’ activity sufficiently for experimental purposes. All the experiments except a few mentioned below were carried out in a room maintained at temperatures between 29 and 32° C.

Ten animals were placed in the chamber and time allowed for the gradient to settle down again. Every 20 min. a record was made of the numbers of animals in the drier half, in the moister half, and on the middle line. After each reading the animals were stirred by means of a bright light until eight or more were attracted into one half. The animals were stirred to the moister and to the drier halves alternately. In one set of experiments, after five readings the chamber was turned through 180° as a control of stimuli arising externally to the chamber. At the same time the hygrometers were interchanged. Another five readings were then taken, making a total of ten readings and 100 positions. In another set of experiments the more carefully controlled experimental method described by Gunn (1937) was used, whereby a total of 200 positions was recorded. The results from the two sets were similar. In both methods each hygrometer records the humidity in each side of the chamber in turn, and the mean of the two humidity differences was the value used.

The results obtained with all stadia of Locusta are given in Table I, and it will be seen that in all cases there is a preference for dry air.

Table I.
graphic
graphic

A comparable set of experiments on immature adults was not done in this series, but similar experiments showed that both male and female immature adults have a preference for dry air (e.g. p. 192).

The strength of the reaction under these circumstances is conveniently expressed as the number of animals in the drier side divided by the number in the wetter side (ratio D/W; “intensity of reaction” of Gunn, 1937). The average percentage on the middle line for all stadia was 11 (minimum 4, maximum 16 for hoppers and adult males, 24 for adult females).

Control experiments carried out in exactly the same way (200 positions) in chambers with uniform humidities, gave an average ratio D/W of 0·98 +; 0·01 (min. 0·75, max. 1·63). The possibility that the results were due to the locusts reacting positively to the sulphuric acid itself rather than to the low humidity, was excluded by experiments in which saturated solutions of such salts as potassium tartrate, potassium chloride, and ammonium nitrate were used instead of the acid (Buxton & Mellanby, 1934). The use of various stirring methods showed that the alternate method described above does not cause a bias simulating a preference. This was confirmed in the experiments where the aktograph was used as an alternative chamber, and no stirring whatever was done (see below). The differences between the ratios D/W for different stadia must be disregarded, since it is impossible to allow for variations in gregarious behaviour with different stadia of different sizes in different sized chambers.

The results show a definite, if not very strong, preference for the drier half of the chamber. The experiments listed in Table I were designed simply in order to establish the existence of a reaction to air humidity. As was pointed out by Gunn & Kennedy (1936), it is not possible to say to what humidity difference the animals are reacting in the alternative chamber, especially when maximum differences of as much as 50 per cent are available. In particular, the results do not reveal whether the locusts have a preferred (“optimum”) region of the humidity range, since such a region may be included in the dry half of the wide range offered in the chamber. In a series of experiments principally on 5th stage hoppers, in alternative chambers in which smaller differences of humidity were provided, a majority of hoppers was always found in the drier half of the chamber, irrespective of where the limiting humidities concerned lay in the humidity range.

The above figures (Table II) of single experiments on 5th stage hoppers show that the dry preference persists even in the driest portion of the humidity range which could be reached experimentally. A ratio D/W of 1 · 4 is actually less than the highest value (1 · 6, see p. 189) obtained in the no-difference controls, but as none of the above experiments gave a negative result the ratio 1 · 4,is probably more significant than the control suggests. As will be seen below, the magnitude of the humidity difference available in the chamber is an important factor in determining the value of D/W. Since the humidity differences are much greater in Table I than in Table II the smaller values of D/W shown in Table II cannot be ascribed simply to the greater dryness of the air.

Table II.
graphic
graphic

Variation in the reaction, even in experiments with roughly equal humidity differences and in about the same region of the humidity range, is, as indicated above, very great. One cause of the variation in the ratio D/W was undoubtedly the wide variation in the level of activity of the locusts (Key, 1936), since it was a property of the experimental technique employed (stirring alternately to wet and dry halves) that the lower the animals’ activity, the nearer the ratio D[Wapproached unity. Experiments with various humidity differences in different parts of the range showed no marked dependence of the ratio D/W on the region of the humidity range, such as Gunn (1937) has found for Porcellio scaber. For instance, when extremes of humidity 18 · 28 per cent R.H. apart (mean 23 per cent) were set up in the chambers, ratios D/Wof 3’6 and 2 · 2 were obtained below 40 per cent R.H., ratios of 4.0 and 2.0 in the middle humidity region, and of 2-1 and 3-2 when both extremes were above 60 per cent R.H. A series of alternative chamber experiments with single immature adult males (illumination 65 metre-candles) did, however, give a higher value of D/W in the moist portion of the range (above 64 per cent R.H.) than in the dry (below 37 per cent R.H.). Although the difference was not statistically significant, it suggests that a larger number of experiments with humidity ranges of equal length might yield a significantly smaller value of D/ W in the moist region than in the dry. On the other hand, Fig. 1 shows that the ratio D/W is roughly correlated with the extreme difference of humidity available. Fig. 1 also shows that when less than 20 per cent R.H. difference is available in the chamber, D/W approaches the values obtained in the control experiments. This 20 per cent difference in the chambers used for 5th stage hoppers represents an average gradient of slightly less than 1 per cent R.H. per cm.

Fig. 1.

Intensity of reaction in relation to difference of humidity available in alternative chambers, 1st and 5th instar hoppers and adult males. Broken line represents highest value obtained in control experiments. A ratio D/W of 1 · 0 is equivalent to no reaction.

Fig. 1.

Intensity of reaction in relation to difference of humidity available in alternative chambers, 1st and 5th instar hoppers and adult males. Broken line represents highest value obtained in control experiments. A ratio D/W of 1 · 0 is equivalent to no reaction.

Effect of low humidity in cages

Experiments were done to find the effect of previous humidity conditions on the humidity preference. Fifth stage hoppers were fed, but kept up to 10 days in cages in which the humidity (measured as described on p. 187) varied between 22 and 32 per cent R.H. One set of hoppers was tested on the 1st and 2nd days, and two sets of hoppers on the 3rd, 6th, 7th, 8th, 9th and 10th days after they had been placed in the dry air cage, making a total of fourteen experiments and 1177 position readings (excluding mid-line readings). The average ratio D/W was 3·8 (1 ·4 −9 ·0), compared with 4 · 8 (2 · 4 − 9·0) for similar animals kept at 45 −80 per cent R.H.

Only two sets of experiments have been done on hoppers kept both in dry cages and without food. On the 4th day of such treatment the hoppers became moribund. With one set of 5th stage hoppers kept in cages at 15 to 18 per cent R.H. a ratio D/W of 2.5 was obtained on the 1st day (hoppers placed in dry cage on previous evening), of 1.6 on the 2nd day, and of 17 on the 3rd day. The second set of hoppers, kept in cages between 24 and 27 per cent R.H., gave a ratio of 14.0 on the 1st, 3.4 on the 2nd, and 1.7 on the 3rd day.

For the technical reason given on p. 190, the changes which the ratio D/W undergoes in the above figures do not necessarily represent a diminution of dry preference, when the level of activity is declining. It can therefore be concluded that keeping hoppers in drier cages for a few days, at least, does not destroy the dry air preference, but the possibility remains that it reduces the intensity of that preference. The detailed results do not demonstrate a reduction of intensity of reaction, but at the same time they do not eliminate such a possibility.

Experiments were carried out with full range chambers (average difference of humidity 60 per cent R.H.) in much cooler conditions in a north light in a room the temperature of which averaged 20 · 0° C. (18 · 6 − 23 · 6° C.). With a total of 429 position readings with 5th stage hoppers, an average ratio D/W of 5 · 4 (2 · 9 − 8 · 7) was obtained, and with a total of 437 readings for immature male adults, a ratio of 6-9 (3-2-8-8). That is, the reaction is still present under conditions of temperature and illumination different from those in which the bulk of the experiments were done.

Fifth stage hoppers were placed singly in the aktograph (Gunn & Kennedy, 1936) in alternately moist and dry air for 1 hour each, under illumination of 5 metrecandles. The basal level of activity varied a great deal from animal to animal, so that the results are purely relative. Fig. 2 shows one case where a very clear reaction was obtained and one which was doubtful. On the whole, the animals were more active in moist air than in dry. In twenty-two cases in which the animal was tested first in dry air and then in moist, there was no change in six cases, and an increase in sixteen cases. In twenty-five cases in which the animal was tested first in moist air and then in dry, there was a decrease of activity in sixteen cases, and no change in nine cases. Experiments in which the activity did not change by more than 50 per cent were included in the neutral class. Moist air tends therefore to make the hoppers more active, and dry air to decrease their activity; they exhibit an hygrokinesis. On the other hand, when the exposure is extended to 23 hours, activity bears no constant relation to the humidity of the air—i.e. the hygrokinesis does not last for long periods.

Fig. 2.

1-hour aktograph records of 5th instar hoppers, a, good reaction; b, doubtful reaction.

Fig. 2.

1-hour aktograph records of 5th instar hoppers, a, good reaction; b, doubtful reaction.

When the aktograph was used as an alternative chamber, that is, when one half of the chamber was kept moist and the other half kept dry, five out of eight records showed that while the locust did not always stay for long periods in one half, the inactive periods occurred predominantly in the drier half during a whole day (Fig. 3). The remaining three records-were ambiguous. Thus the continuance of the hygrokinesis of Locusta, unlike that of Porcellio, requires occasional exposures to a different humidity from that preferred.

Fig. 3.

23-hour record of a 5th instar hopper in the aktograph used as an alternative chamber. The continuous fall is due to distillation of water from the moist to the dry side. Animal in moist side during upper halves of strokes, in dry side during lower halves of strokes.

Fig. 3.

23-hour record of a 5th instar hopper in the aktograph used as an alternative chamber. The continuous fall is due to distillation of water from the moist to the dry side. Animal in moist side during upper halves of strokes, in dry side during lower halves of strokes.

Single individuals were observed in alternative chambers and their tracks drawn (Fig. 4). There was no evidence of a clear-cut avoiding reaction as a component of the humidity reaction. Examination of Fig. 3 shows that on many excursions away from the drier end of the aktograph, the locust does in fact return into the dry half without going right into the moist end. This was also observed in the alternative chamber, but the reaction by which it is effected is difficult to allot to any of the recognized categories of behaviour (Fraenkel, 1931). When passing from the drier half towards the moist end of the chamber, the animal was often seen to turn aside, and veer round until it returned into the dry region again. Sometimes, after turning, it walked along or parallel to the middle line ; with some individuals this, perhaps hygrophobotactic, reaction occurred very often, while with others it did not occur at all. Progress from the moist to the dry was, on the contrary, usually uninterrupted.

Fig. 4.

Path followed for about 8 min. by the head of an adult male Locusta in an alternative chamber. Broken line represents boundary between the two sets of humidity controlling dishes, s., start; f., finish; tt., turning reactions. It will be noticed that in the middle of the chamber the animal usually passes uninterruptedly from the moist into the dry side, but executes more or less sharp turns when moving from dry to moist.

Fig. 4.

Path followed for about 8 min. by the head of an adult male Locusta in an alternative chamber. Broken line represents boundary between the two sets of humidity controlling dishes, s., start; f., finish; tt., turning reactions. It will be noticed that in the middle of the chamber the animal usually passes uninterruptedly from the moist into the dry side, but executes more or less sharp turns when moving from dry to moist.

Attempts were made by various means to evoke a turning reaction through asymmetrical stimulation of single individuals. Small dishes of water and sulphuric acid were placed under a perforated floor under the two sides of the animal. Currents of moist and dry air were allowed to play on opposite parts such as antennae, palps, thoracic spiracles, and abdomen. No evidence of orientation under the influence of these stimuli has so far been obtained.

It has been stated above that the least gradient in an alternative chamber to which animals show a good reaction is about 1 per cent R.H. per cm. The actual humidity difference between symmetrical humidity receptors (if they exist) is therefore very small. This fact, taken in conjunction with the negative results from the experiments where much greater humidity differences were applied to opposite sides of the animal (p. 195), makes it very unlikely that the reaction is tropotactic. The reaction appears rather to resemble the behaviour of planarians (Dendro-coelum lacteum) in a gradient of light intensity (Ullyott, 1936). On passing into a region of higher light intensity, these animals tend to change direction more frequently than before, but the rate of change of direction falls off again owing to adaptation if no further increase of intensity is encountered. This kind of behaviour leads the planarians to collect in the darker end of the gradient. The turning reaction of the locust has not been investigated in the light of these results but it will be noted that the rate of change of direction is greatest where the humidity gradient is steepest (Fig. 4) and that adaptation occurs in the hygrokinesis (p. 192).

With regard to the work of other authors, Rubtzov (1935) described an experiment in which the preferred temperature of a locust, Chorthippus albomarginatus, was raised by drying the air at the cool end of the temperature gradient and moistening the air at the warm end. The absence of any measurements of the actual humidities in this experiment makes reliable intepretation impossible, and the possibility that in these very experiments the locusts are actually preferring drier air is not excluded. Key (1936) conducted experiments on the locomotory activity of Locusta in relation to air humidity by methods very different from mine. His animals were reared either in very wet or in very dry conditions, while mine were reared at 50-80 per cent R.H. In the only experiments which appear capable of comparison with mine, namely some of those in which the animals were reared in very wet conditions, he found that 3rd, 4th and 5th stage hoppers were more active in dry air than in wet, the opposite of my result. No explanation of this contradiction is apparent to me, but it may be said that when analysed in the same way as his results, my aktograph results show a much larger difference between the activity levels in the two humidity conditions than he obtains.

The fact that in all the experiments described above the locusts show a preference for dry air, makes it almost impossible to draw conclusions from them having any bearing upon the available field observations. The condition of high temperature combined with low illumination may seldom be encountered by locusts in the field, but the experiments at 20 ° C. in daylight showed an even stronger preference for dry air, so that the reaction is not dependent on the former combination of factors. Yet Rubtzov (1935), for instance, writes: “Although the Acrididae are generally xerophytic animals, their second and third stage larvae invariably move towards higher humidity, and congregate there in dense bands.” Hamilton (1936) has shown that the optimum conditions for the development of Locusta are between 60 and 75 per cent R.H. Lean (1931) found no breeding in the field below 40 or above 80 per cent R.H. Thus the low humidity preferred by locusts in my experiments is not the optimum humidity for development, maturation, and breeding.

With regard to the migration and dispersal of swarms, Lean (1931) observed during an invasion of Nigeria by Locusta that the swarms were to be seen only between the 40 and 85 per cent lines of equal relative humidity throughout the infestation. Similarly, Smee (1936) found that the circling flights of swarms of the Red Locust (Nomadacris septemfasciata) tended to change into flights in one direction, which carried the swarms out of the locality in question, whenever the temperature rose above 27° C. and the humidity fell below 60 per cent R.H. It appears from his observations that humidity is a rather more important influence than temperature.

Immediate application of such results as those obtained in the course of this work could not be expected. The prevention of hopping and flying, the low illumination, and the very small distances involved in the experiments, render them inapplicable to the problems of locust distribution and migration. The reaction is in any case weak compared with that of Porcellio (Gunn, 1937), and the ease with which it may be overcome by an opposing light stimulus (e.g. by the stirring in the alternative chamber experiments) emphasizes this point. According to Fraenkel (1929) and Strelnikov (1936), temperature is the dominating factor in the diurnal behaviour of locusts. Humidity is ecologically inseparable from temperature, so that the reaction may play at any rate some part in diurnal activity.

The apparent contradiction between the results of this work and the field observations will only be resolved by a careful study of physical factors in relation to locusts in the field.

This work was carried out with the aid of a grant from the Imperial Institute of Entomology in the Department of Zoology, University of Birmingham. I have to thank the members of that Department for the interested and helpful atmosphere which I found there; in particular, my thanks are due to Prof. H. Munro Fox for his hospitality and for a considerable amount of apparatus, and to Dr D. L. Gunn for unfailing help and advice. Acknowledgements are also due to Dr B. P. Uvarov for his constant encouragement and for originally suggesting the work, and to Dr A. G. Hamilton for supplies of locusts and for technical advice on their culture.

Bodenheimer
,
F. S.
(
1930
).
Studien zur Epidemiologie, Oekologie und Phytiologie der afrikanischen Wanderheuschrecke, Schistocerca gregaria Fonk
.
Berlin
.
Buxton
,
P. A.
&
Mellanby
,
K.
(
1934
).
Bull. ent. Res
.
25
,
171
5
.
Fraenkel
,
G.
(
1929
).
Biol. Zbl
.
49
,
675
80
.
Fraenkel
,
G.
(
1931
).
Biol. Rev
.
6
,
37
87
.
Gunn
,
D. L.
(
1937
).
J. exp. Biol
.
14
,
178
86
.
Gunn
,
D. L.
&
Kennedy
,
J. S.
(
1936
).
J. exp. Biol
.
13
,
450
9
.
Hamilton
,
A. G.
(
1936
).
Trans, ent. Soc. Lond
.
85
,
1
60
.
Key
,
K. H. L.
(
1936
).
Bull. ent. Res
.
27
,
399
422
.
Lean
,
O. B.
(
1931
).
Bull. ent. Res
.
22
,
551
69
.
Rubtzov
,
I. A.
(
1935
).
Zasch. Rast. Vredit
.
3
,
33
8
.
Smee
,
C.
(
1936
).
Bull. ent. Res
.
27
,
15
35
.
Strelnikov
,
I. D.
(
1936
).
Trav. Inst. zool. Acad. Sci., U.R.S.S
.,
2
,
637
733
.
Ullyott
,
P.
(
1936
).
J. exp. Biol
.
13
,
265
78
.