The Behaviour of Wireworms in Response to Chemical Stimulation

If wireworms are ever to be controlled by trapping it is necessary first to discover how they find their normal food under natural conditions. The handicap created by the scarcity of laboratory experiment on the food finding and feeding behaviour of wireworms (Agriotes sputator, obscwrus, Hneatus) must have been obvious to all those who have tried to devise a practicable field method of control by trapping. Without such basic laboratory observation properly co-ordinated with field experiment there is little prospect of knowing whether or not baiting can be a practical technique of control, and all attempts to elaborate an efficient method are bound to be empirical. It was for this reason that the present work was undertaken.

The periodic movements of wireworms in soil are of course vitally important in relation to the problem of food finding. It has long been realized that wireworms tend generally to move downwards in the soil during the late autumn and to some extent during periods of summer heat and drought and upwards again in spring and late summer. It has recently been shown that such movements are probably based on orthokinesis, and on a ‘shock reaction’ to temperature extremes and to humidities below saturation and are not brought about by any response to gravity—such a sense being completely lacking to the wireworms when burrowing in soil (Lees, 1943 a, b; Falconer, 1945 a).

It is difficult to evaluate field observations on the efficiency with which wireworms find their food in the soil. In grassland there is clearly no problem. Here most individuals probably live in semi-permanent systems of burrows and remain fairly near the surface. Those that go down deeper in winter could hardly fail on rising in spring from the lower levels of the soil, to strike grass roots on which to feed. But under arable conditions things may be very different. While wireworms here, too, form their semi-permanent systems of burrows (Lees, 19436), used once they have found their food, they certainly tend to descend more deeply and uniformly in winter, and during this period their burrows will soon get destroyed by cultivation and the effects of weather. Obviously, if when they rise again in spring a large proportion of the surface is covered with crop plants or if there are well-developed root systems, the wireworms will have little difficulty, merely by random burrowing in the soil, in finding something to eat. But though under such conditions they may find food in the form of fine rootlets, this, since the larvae beneath the soil lack all guidance from light and gravity, might aid them little in finding the succulent central parts of the plant. Now it is clear that the efficiency of food finding of wireworms does vary with environmental factors such as soil texture and method of cultivation. Nevertheless, one cannot fail to be impressed by the speed and apparent certainty with which, at any rate in some instances, even quite small plants can be located in spring. Erratic as the results of experiments with trap crops are apt to be, they do provide some evidence of considerable powers of orientation to food, at any rate over short distances (Miles & Petherbridge, 1927; Ladell, 1938; Petherbridge, 1938), and the records of the rapid and very high infestation of potato tubers by reasonably small populations of wireworms {Bull. Minist. Agrie, no. 128, 1944, pp. 44-5) also point to this conclusion. It seemed possible, therefore, that a chemical method of orientation might be involved. Preliminary experiments by one of the authors (W.H.T.) had been successful in showing that responses of wireworms to plant extracts and to individual chemical substances could be studied conveniently in the laboratory. The object of the present paper is to give an account of that part of the work which was primarily directed to answering the question—‘How do wireworms find their food in the soil?’ The greater part of the laboratory work here described has been carried out by two of us (A. C.C. and J.H. D.). Further papers on the relation between activity and chemical constitution (A.C.C. and J.H.D.) and on the sensory equipment of the wireworm (A.C.C.) are in preparation.

Because of the difficulty of keeping large stocks in healthy condition in the laboratory for any considerable period we decided to rely on material freshly collected from the field. During the first 2 years it was possible to obtain large samples from farmers and from the various wireworms survey teams of the Ministry of Agriculture. Later, owing to the development of mechanical methods of sorting soil samples, the latter source dried up almost completely, and it was soon found that the careless handling and delays inseparable from the former method might involve such a large mortality or, at best, loss of condition, that it was best avoided. Accordingly for the greater part of the time high wireworm infestations in the neighbouring districts of Cambridgeshire were kept under regular observation, and our material directly collected by ourselves as and when soil and cropping conditions made the extraction of large numbers feasible. For help in finding suitable infestations we are indebted to Mr F. R. Petherbridge, Cambridge School of Agriculture, and his assistants of the advisory staff; to the officers of the Cambridge War Agricultural Executive Committee; and among farmers particularly to Mr W. Jackson of Hill Farm, Upware, whose success in overcoming by good cultivation and management a high wireworm infestation on what was previously semi-derelict fenland, has been remarkable. We are also greatly indebted to Mr Percy Hardwick of Fulboum, to Mr G. W. J. Burden of Hardwick, and to Mr Hawkes of Barton, as well as to Mr Jackson, for providing invaluable facilities for field experiment, and for much co-operation and forbearance without which the work could not have been completed.

Like all other workers with wireworms we have had to accept the present impossibility of distinguishing between A. lineatus and A. obscurus in the larval stages. Thus our material of these two species may have been either one or other or mixed populations of both. Whenever we have been able to obtain numbers of A. sputator we have, as is made clear below, used this as well. Our results have given no indication of there being any substantial difference in behaviour between sputator and lineatus-obscurus populations. We therefore assume for our present purpose that the three species are identical in food-finding responses. There does not at present seem to be any definite indication from the field against this assumption.

Many previous workers have found that populations of wireworms show two moderately well-defined non-feeding periods each year, one during the winter and the other in the early summer. Conversely, there are peaks of feeding activity in spring and autumn. The fasting periods are no doubt in part an expression of the vertical soil movements which are governed by temperature and humidity changes (Lees, 1943 a, b; Falconer, 1945 a), but it has been shown by the study of individual wireworms (Evans, 1944) that each moult is followed by a feeding phase of a few weeks duration followed by a slightly longer fasting phase. Larger wireworms moult less frequently than smaller and tend to remain longer in the fasting phases. Thus the feeding behaviour of a population is largely the expression of the heterogeneous physiological rhythms of its constituents. As it is more convenient to use the larger wireworms for experiment, the dependence of feeding period on the size raises a difficulty. It was found that while the biennial rhythm may be modified by transference from field to laboratory conditions, the fasting periods cannot be eliminated thus. The seasonal variation in the percentage of wireworms giving the orientating response in routine tests throughout 1943 and 1944 is shown in Text fig. 2. This shows how far active and inactive periods correspond to the feeding and non-feeding periods of Evans (1944). It also shows how during certain times of the year experiments may be impossible owing to the high proportion of non-feeding individuals in a population. Raising the temperature of the stocks from 16 to 23° C. or subjecting them to a preliminary cold shock at 5 ° C. for 2 days did not significantly increase the number responding. All batches of wireworms used for orientation experiments were tested against a standard ‘bait’ of 2% sucrose to determine their activity. Later, when a technique for studying the behaviour of single wireworms in the plate apparatus was perfected (see below), these difficulties were avoided by selecting only active wireworms. All the larvae used in these experiments were from i’5 to 2-0 cm. long, representing the later instars (Salt & Hollick, 1944).

All experiments with the apparatus here described were performed in a constant temperature room at either x6 or 23° C.

To test the responses to airborne odour an ordinary glass Y-tube olfactometer 5 cm. in internal diameter and filled with glass beads to provide the necessary tactile stimulation was first employed. Wireworms showed no response to odours in this apparatus. The apparatus, however, is inefficient for a relatively slow-moving soil-dwelling animal. The area of contact between the two air streams to be tested is small, and the light beam necessary to ensure that the animals travel up the tube to the junction introduces an unnatural factor into the situation. To avoid some of these difficulties a choice chamber was devised.

This second apparatus was a modification of that described by Wigglesworth (1941). Into a deep Petri dish across which a watertight partition had been fixed were introduced, on opposite sides of the partition, equal volumes of distilled water and the solution to be tested. The surfaces of the liquids on the two sides of the partition were at the same level. A metal cylinder, 1 in. deep, across one end of which a piece of fine brass gauze had been fixed, was now placed, gauze downwards, into the Petri dish. Supports were arranged so that the gauze was a few millimetres above the surface of the liquid. Beads were now placed in the space above the gauze, and a number of wireworms placed on top of the beads in the centre of the cylinder. The whole was now lowered into a deep, blacked-out water-bath (maintained at 2 ° C. above air temperature) so that the bottom of the Petri dish was supported below the surface of the warmed water. The space above the water in the waterbath was saturated with water vapour. Vapour from the distilled water and the odorous substance rises up through the gauze and the beads, and the wireworms above the gauze would show a response to the odour concerned by aggregating on one side or the other of the partition. The numbers on either side were counted after 24 hr. In this apparatus, likewise, no response to odours was observed.

These negative results might have been due to the unnatural conditions in the two preceding types of apparatus preventing the wireworms responding normally to chemical substances borne in the air. A third apparatus was therefore designed in which it was hoped that conditions would be more natural.

The plan and elevation of this apparatus, which was constructed of metal, are shown in the accompanying diagram (Text-fig. 3). The damp sand in the central space just covers partition B, so that an unbroken surface of sand stretches from A to C. The glass partition B¹ which divides the air space into two compartments comes down to within about 5 mm. of the surface of the sand, so that a wireworm can easily pass under it. The glass lid is made airtight with a rubber gasket. The ‘bait’ solution is either placed in one of the side troughs divided off by partition A or C (the other containing an equal volume of distilled water) or air is bubbled through a ‘bait’ bottle before entering the inlet tubes on one side as was done with the Y-tube apparatus, or finally the bait may be placed in both positions. Wireworms entering the apparatus at one end below the central glass partition encounter the junction between the two streams of air, and after wandering about on the surface of the damp sand, eventually burrow down into it. A response to the odour concerned would be shown by their aggregation on one side or the other of the central partition (§ IV (ii) (a)).

The responses of wireworms in sand or soil were investigated by two methods. The first made use of a metal box 13 × 19 cm. and 15 cm. deep with a movable metal or glass partition dividing it lengthwise into two (Text-fig. 4). Washed sand to a depth of 3-25 cm. (800 ml.) was placed in the bottom. The sand on one side was moistened by thoroughly mixing with 5 ml. of water, that on the other with the same quantity of solution to be tested. To start the experiment the partition was removed and the wireworms placed in the middle region of the sand surface. At the end of the experiment the numbers in the ‘baited’ half and the unbaited half of the apparatus were counted. This choice chamber thus tests the ability of the insects to orientate in sand or soil in solutions of a given substance. All ‘aggregation experiments’ were performed at 16 ° C., except when attempting to break dormancy when they were performed at 23 ° C. (see above). There was no observable difference in responses given at the two temperatures.

The second method makes use of the fact that wireworms will bite filter paper soaked in solutions of certain food substances and their bite marks can easily be counted, thus providing, when compared with a control set of filter papers moistened with water, a numerical measure of the ‘palatability ‘of a given substance. Indi vidual wireworms may vary considerably in the number of bites given on active substances; for this reason they were used in groups of 20. The amount of biting may be the index not only of palatability but of feeding activity. The latter was therefore standardized by measuring the amount of biting elicited by a standard solution (2% glucose) from samples of each group of wireworms collected. Most groups gave 150-600 (average 350) bites per 20 wireworms in 24 hr., those biting beyond these limits being rejected for subsequent tests. This represents the usual amount of biting and degree of variation of most of the groups collected (see discussion on Text-fig. 6). Wireworms bite the filter papers impregnated with an active substance and also the controls containing water. The amount of biting on the latter is rather variable and bears no relation to the amount falling on the baits, but does not exceed about 70, that is, 20 % of the total number of bites given in positive tests. This represents the residuum of biting not chemically stimulated. Because of this variability the controls were ignored when estimating ‘palatability’, the total number of bites per 20 wireworms in 24 hr. being taken as the index of this factor. As a qualitative test the method was also useful in determining whether some of the less palatable substances were palatable or not: the bites on filter papers with palatable substances are characteristically close together (Pl. 9). On the controls they are distributed evenly over the filter paper.

A large number of experiments with baited filter papers were carried out in an apparatus consisting of two sheets of plate glass a foot square separated by a marginal rubber gasket 5 mm. thick and held in a wooden frame. The lower glass sheet is covered with a layer of paraffin wax 1 mm. thick in which are eight circular cavities, 5 cm. in diameter, arranged in a ring. A piece of filter paper (5 cm. in diameter) is fitted into each of the eight cavities. Four alternate filter papers are moistened with five drops (0-25 ml.) of the solution to be tested; the other four with the same quantity of water. The apparatus is then filled with glass beads in a saturated or nearly saturated atmosphere, 20 wireworms introduced, and the top sheet put into position. After 24 hr. in darkness the filter papers are removed and the number of bites on each counted under a binocular microscope.

It was found that baited filter papers could be used equally well under damp sand in the bottom of the choice chamber box as shown in Text-fig. 4, and this has the advantage of avoiding any difficulty due to an unsaturated atmosphere. This was the method generally adopted in later experiments. The sand was passed through sieves, and only grains between 0.1 and 0-2 mm. in diameter were used in this and other apparatus described in this paper. The sand was 14 ·3% saturated with water. The value for saturation was obtained by adding water to the sand in a measuring cylinder until free water just appeared at the surface (Lees, 19436). The sand was ¹4’3% saturated when 5 c.c. water were added to 400 c.c. sand. Both biting and orientation took place best with this amount of moisture. This is shown in Table 1. In 28 ·6% saturated sand there was practically no biting and the wireworms were 100 sluggish to move. The 5 · 7% saturated sand, on the other hand, was much 100 dry; the wireworms gathered on the wet filter papers or huddled together in the comers of the apparatus. But biting does occur even at this degree of saturation: in one experiment in which the wireworms gathered on a sucrose filter paper, this was closely bitten. In the 14 · 3 % saturated sand the wireworms exhibit both the normal orientation and biting responses. All ‘biting experiments’ were performed at 23 ° C. The biting response is exhibited at 16 ° C., but the number of bites given is less.

The mechanism of orientation of wireworms to solutions in soil was examined by means of an apparatus similar to that used by Lees (1943 b) and Falconer (1945 a) for observing their tracks in sand or soil.

This apparatus consists of a glass plate to which are cemented glass strips 12 cm. long and 2mm. deep forming a series of square compartments. Each compartment is marked into two by a line scratched across the middle. When testing responses in soil, each compartment is filled with sieved air-dried fen soil which is rolled smooth and level with the top of the glass strips. The soil may then be impregnated with suitable amounts of solutions of various substances by spraying from a fine pipette either the whole surface of each compartment or, if a boundary is desired, half of its surface with the solution and half with the same amount of water. After the liquids have soaked in, the surface is rolled again and a wireworm placed in the required position ; each compartment is then covered with a glass plate. It can be seen that the animals burrow by pushing aside the fine soil and by picking up the larger crumbs and grains in their mandibles and placing them on one side. (E. T. Burtt unpublished MS.) The tracks thus made are traced after a suitable lapse of time, usually two hours. When testing responses in sand the procedure is similar, but the water and solution, respectively, are mixed with the sand before the latter is placed in the compartment. The tracks made in sand may be most readily seen by illuminating the apparatus from below. The experiments were performed in darkness at 16⁰ C. The sand used consisted of particles between 1 and 2 mm. in diameter. The soil was the fraction that passed through sieves with 8 meshes to the inch, but was retained in sieves with 30 meshes to the inch.

That wireworms are able to find growing wheat either by moving through the soil or by coming to the surface and moving over it was shown by the following experiment (cf. Falconer, 1945 b). Wheat was grown in a row in clean sand in one section (C) of the container shown in Fig. 1. Wireworms were released in section B which was separated from A and C by two sets of partitions, one in the air reaching to the surface of the sand and the other reaching from the bottom of the air partition to the bottom of the container. When the wireworms had burrowed beneath the sand, in different experiments either all the partitions or only those in the sand or in the air were removed. The sand in the three sections was sieved after 48 hr. and the wireworms counted. The results (Table 2) show that wireworms can move both through the sand and over its surface, and that, in either event, more collect in section C containing the wheat than in section A containing none.

It was originally intended to investigate how far the apparent food preferences exhibited by wireworms in the field are the expression of selective response to differences in chemical composition. But the field evidence for food-plant preference is very confusing, since all that is usually recorded is the number or proportion of plants destroyed or showing injury (cf. Subldew, 1934). There is thus no means of allowing accurately for the fact that some resistant crops (e.g. potato) can survive a considerable attack without showing much damage above ground, whereas others (e.g. freshly germinated cereal) may be destroyed by the first bite. Evans (1944) showed flax to be physiologically unsatisfactory as a nutritive material for the wireworm, but filter papers, impregnated with the juice of its roots were bitten (§ IV (i) (a)), and there seems to be no real evidence for assuming that this or other so-called ‘resistant’ crops are in any degree unattractive or unpalatable to wireworms. Most so-called resistant crops are probably not repellent or immune from attack to any extent but are merely more able to recover from it when it occurs. For this reason and jlso because of the complexity of the chemical problem we did not proceed with this part of the work.

The biting response of fed wireworms is elicited by the juice of potato tubers, carrot and sugar-beet tap roots and wheat and flax roots, and by an aqueous extract of wheat bran. The extract was placed on filter papers in the biting apparatus described above. The first problem was to discover the active chemical substance in these extracts.

The activity of substances causing either of the responses is defined as follows : If i g. of a substance is dissolved in x ml. of water to reach the threshold, then the activity of that substance is defined as log₁₀x. For example, a substance for which the threshold is a 1 % solution has an activity of 2. The threshold is defined as the lowest concentration which causes any response.

Among other substances, the sugars glucose, fructose, and sucrose (Table 3) are active in concentrations at which they are present in the plants. For instance, with fed wireworms the palatability of sucrose falls from the maximum to zero between 0-5 and 0-2%, that of glucose between 1-26 and 0-5% (Text-fig. 6). Expressed potato juice, which contains about 1% glucose and 0 · 1% sucrose, is active but becomes inactive when diluted 10 times. Sugar-beet juice, which contains about 18% sugars, is active when diluted 10 but not 100 times. An extract of bran with 10 times its weight of water contains 0 · 4-0 · 6% sugars. It is clearly the sugars which elicit the biting response to these plant extracts.

The effect of pH on biting was tested as follows. 2 % glucose solutions were buffered at different pH values from 0 to 14. The buffers themselves were previously tested and found inactive. The following buffers were used to obtain the pH values

The wireworms were starved for 1 week before the experiment. When the total number of bites given by 40 wireworms on filter papers soaked in glucose solution is plotted against the pH of the solution, the curve shown in Text-fig. 5 is obtained. Maximum biting occurs between pH 6 and 8, and biting decreases as the solutions become more acid or more alkaline. The number of bites on the controls varied between o and 40. Most plant juices have a pH of 5-7.

A list of the chemical substances (2% aqueous solutions unless otherwise indicated) tested for the biting response is given in Table 3. Two or more experiments were performed with each substance. The wireworms were starved for 1 week. The average number of bites elicited by most of the active substances was about 350 per 20 wireworms in 24 hr., with a possible variation with each substance from 150 to 600. The substances marked with an asterisk were less palatable. Some of the latter, viz. arabinose, dextrin, hen’s egg albumen, horse-serum albumen, haemoglobin, tannic acid and sodium tannate, sometimes elicited as many bites as members of the most palatable group, but the number elicited was usually between 50 and 200, and was occasionally no greater than that on the controls. Mannose, mannitol and gluconic acid only occasionally elicited more bites than the controls, up to 120 per 20 wireworms in 24 hr. There is, however, no doubt that these three substances are feebly active: this is indicated not only by the number of bites but by the characteristic close biting on the filter papers containing them as compared with the controls. These ‘less palatable’ substances did not elicit a greater number of bites at higher concentrations, e.g. in 5% solutions. Arabinose, glucose, mannitol, triolein, sodium stearate, casein, peptone and tannic acid were tested with A. sputator as well as A. obscurus-lineatus, with which all the substances were tested. No differences in biting behaviour were observed between wireworms of the different species.

The compounds which cause biting are all members of the three major food groups, carbohydrates, fats and proteins, and are tabulated in this grouping in Table 3. Tannic acid is an exception and gives a weak response.

In the carbohydrate group not only are the common plant sugars—glucose, fructose and sucrose—highly active but also other sugars ranging from mono-to tetrasaccharides and dextrin. Solutions of starch and inulin do not cause biting. In sorbitol, mannitol and gluconic acid the aldehyde group has been altered and the ring structure destroyed but the compounds remain active. This suggests that the polyhydric alcohol grouping is responsible for the activity as in the case of the sweet taste in man. In the biting response there is, relative to human taste, a shift towards larger numbers of alcohol groups as the wireworm is sensitive to dextrin but not to glycol or glycerol. This resemblance to the sweet taste is confined to polyhydric alcohols, for glycine and saccharin cause no response.

Fats have been tested both as emulsions stabilized with sodium oleate or cetyl trimethyl ammonium bromide and by depositing weighed amounts on filter papers from a solution in ether. The latter method is preferred as it is less variable, but the compounds shown in the table have been tested as 1 % emulsions except where the contrary is stated. Triolein is the only pure fat found to be active, but as linseed oil contains less than 5% of its fatty acids as oleic acid (Cocchinaras, 1932) the glycerides of either of its other major component acids, linolic and linolenic, may be active. Tristearin and tripalmitin are not active, neither is glycerol, but these two fatty acids are active as sodium salts. Wheat-germ oil is active as a 6% emulsion but not at 2%. It contains 28% of its fatty acids as oleic (Jamieson & Baughman, 1932).

The proteins so far found to be active are of animal origin, while the plant proteins tested are inactive. Partially broken-down proteins may be active though the parent proteins were not, but none of the amino acids or mixtures of amino acids tested have proved active.

The activity of all the substances shown in Table 3 is at least 1-7 for wireworms starved for 1 week, i.e. 2% solutions are active. For fed wireworms the activity of sucrose was 2-7 and that of peptone 2 to the nearest o-2 log unit, while the threshold for triolein was between 17 and 15 mg. per filter paper (5 cm. diam.) or between 1 and o-i % emulsion. The activity of glucose was more precisely determined with fed wireworms of the species A. obscurus-lineatus, all collected from the same place and of standard feeding activity. They had been fed on potato slices up to the beginning of the experiment. The results are shown in Text-fig. 6, in which the activity corresponding to each value for the concentration of glucose is shown on the abscissa. The average number of bites elicited per 20 wireworms in 24 hr. decreased from a maximum of about 350 when the concentration fell below 1-26% and reached a value no greater than that on the control filter papers at 0-5%. The variation in the number of bites is considerable at the higher concentrations, but decreases with concentrations below 1-26%. The same wireworms were now starved and their sensitivity to glucose measured after different intervals. Time is measured from the last feed to the beginning of each test. As shown in Text-fig. 7, up to 7 days the animals become more sensitive to glucose solutions as starvation increases. Starvation beyond 7 days has no further effect on the threshold. The lowest concentration persistently eliciting a response is 0-2% corresponding to an activity of 2-7, although definite biting occasionally occurs at 0-126% (activity 2-9). It seems unlikely that the sense organs of fed wireworms were adapted at the beginning of the experiment to the sugars in the potato on which they were fed, and that the fall in the threshold with starvation corresponded to the disappearance of adaptation. The time taken seems altogether 100 great for such a process. A more probable explanation is that the limit of sensitivity of the receptors is reached after 7 days’ starvation, this not being realized in the behaviour of fed wireworms because some central inhibiting factor, whose effect decreases with increasing starvation, intervenes between stimulus and response. Starvation has a similar effect on the behaviour of various other insects, e.g. the red admiral butterfly, Pyraméis atalanta (Minnich, 1922), blowflies of the genus Calliphora (Minnich, 1929, 1931 ; Haslinger, 1935) and the honey bee (Minnich, 1932; von Frisch, 1934).

The responses of wireworms to airborne odours were tested in the olfactometer described above (Fig. 3). No response was observed to air which had been bubbled through mashed up potatoes or carrots. A petrol (below 40° C.) extract of potato pulp was allowed to evaporate on cotton-wool. After the petrol ether had been removed there was no response to air passed through the cotton-wool. Wheat was grown in a shallow dish of damp sand till a mat of roots was formed. This was placed for 48 hr. while still growing on a pad of 10 g. of cotton-wool containing 2 g. of lard evenly spread by evaporation from a solvent. The lard absorbed volatile compounds from the roots and the pad had a strong vanillin-like smell. Air saturated with water at room temperature was passed through the cotton-wool in a flask heated to ioo° C. in a water-bath, and cooled before passing through the olfactometer. The wireworms did not respond, though sufficient volatile compounds were carried over for the human nose to distinguish between the two streams of air emerging.

(6) Responses to dissolved substances. When the aqueous extract or juice of flax seedlings, wheat bran, sugar beet, mangold or carrot tap roots, or potato tubers was mixed with the sand on one side of the choice chamber described above (§ II) and the same amount of water with that on the other side, wireworms of all the species used orientate to the plant extract.

In the original orientation tests, any significant difference from 50 % was decided by inspection ; no statistical analysis was carried out. After a sufficient number of tests had been made, a statistical analysis was carried out on the whole of the results available. The percentage of wireworms found on the bait side was calculated for each of the three batches of about 100 wireworms used in testing one substance. The accompanying frequency diagrams (Text-fig. 8) show the numbers of experimental results occurring between limits of 5 % for the substances classed as positive and for the remaining tests, which are here called neutral. Tests may be neutral for two reasons: either the test solution or the wireworms may be inactive. The symmetry of the neutral diagram shows that the positive results have not been obtained by arbitrary selection of results greatly exceeding 50%. It also suggests that very few or none of the substances tested give negative (repellent) results, although many are positive and fewer neutral. The normal distribution cannot rigidly be applied as this case is limited in two directions, viz. o and 100%. In the neutral diagram, at least, the tails are far enough from o and 100% for an estimation of the standard deviation, considering the distribution to be normal, to be used as a test of the significance of any deviation from 50%. In fact, the third and fourth cumulants g_± and g₂ for the neutral diagram are 0-013 and 0-309, and their standard errors are ±0-227 ^and ±0’452 respectively; for the positive diagram and₂ are 0-400 and 0-381, and the standard errors are ± 0-168 and ± 0-333. This means that the neutral diagram is not significantly different from the normal distribution, and the difference of the positive diagram is only just significant (Fisher, 1934). For the neutral diagram, the standard deviation of one result in this test is 7-8 %, and for the mean of three results the standard error is 4-5 %. If the difference from 50% of the mean of three results is greater than twice the standard error the solution used as bait is considered to be active. The standard deviation for the positive diagram is 8’2%.

When the concentration of an active substance is increased above the threshold there seems to be no increase in the number of wireworms on the bait side. Neither is there any difference in the number collecting on the bait side in different active substances each above the threshold. These observations are borne out by the positive frequency distribution which is based on the response (as defined above) to many different substances and to concentrations of single active substances varying by as much as a factor of io⁸; yet the standard deviation is not significantly different from that of the neutral curve. Any active substance above the threshold causes 64% of the wireworms to collect on the bait side with random variations corresponding in magnitude to those which occur when no active substance is present. Another test of activity is whether wireworms respond to a substance in the plate apparatus (§ III (d)). Tests with individual wireworms in this apparatus show that many of them are inactive and that active individuals may wander about on both sides of the boundary between substance and water. These facts may explain the low proportion collecting on the bait side in the choice chamber.

Orientation is brought about by sugars (Table 4), their activity being between 2 and 3 for fed wireworms, and, contrary to what was stated in our letter to Nature (Thorpe, Crombie, Hill & Darrah, 1945), also by other bitten substances. The expressed juices of potato, mangold and carrot each have an activity of 6 for fed wireworms. A more detailed examination was made of potato juice. Potatoes were ground and the juice pressed out, spun to remove starch grains and filtered through kieselguhr. The activity of this juice was 6. The sugars in the juice, found to be 1 %, would only account for an activity of 0-1 on a logarithmic scale, so that some other active substance or substances must have been present. After boiling and filtering, followed by boiling with 2-5% of charcoal, a colourless protein-free solution of activity 6 was obtained. The addition of two volumes of alcohol gave a copious inactive precipitate, and a third volume precipitated colourless rhombic active crystals. These proved to be asparagine, the identification of which was confirmed by analysis of the copper salt. (Cu(C₄H₇O₃N₂)₂ requires Cu 19-6%, N 17-2%; found Cu 19-2%, N 17-0%.) Asparagine has an activity of 9-11 and glutamine the same activity. The distribution of asparagine and glutamine in potatoes has been examined by Neuberger & Sanger (1942). The means of their results in six varieties are respectively 0-26 and 0-21 % of the fresh weight. This range would account for our observed activity of the juice, whether the activities of asparagine and glutamine are additive or not.

The substances causing orientation are shown in Table 4. All the activities shown in this table have been determined with wireworms starved for 2 days using sand in the choice chamber. Activities were measured to the nearest 2 log units, i.e. an activity of 9 means that activity was greater than 9 but less than 11. Compounds which are not active at a concentration of 1 % are considered to be completely inactive for practical purposes, and in the table their activity is specified as ‘none ‘instead of <2. Substances causing orientation consist of a group of sugars all of activity 2-3 and peptone, triolein and tannic acid each of activity 2-3, and of a group of acids and amides with higher activities. In the latter group are several dibasic acids and amines widely distributed in plants. The amides of the lower fatty acids are active, but the acids themselves and ammonia are not. It is interesting that amides are the active compounds in buffalo dung which attract the Buffalo Fly Lyperosia exigua. (Krijgsman and Windred, 1933).

That pH has no effect between 4 and 8 has been shown by diluting the solution to be tested with inactive buffer solutions. No difference has been found between acids and their sodium salts. At such low concentrations as are used both the acid and the salt are so nearly completely ionized that the anion must be the active agent. At a concentration of i in io⁸ acetic acid is about 99 % dissociated and has a pH close to 7.

Glucose, sucrose, peptone, triolein, asparagine, aspartic acid and acetamide were tested with A. sputator as well as A. lineatus-obscurus with which all the substances were tested. No difference in behaviour was observed between the different species of wireworms. Nor were activities different at 16 and 23⁰ C.

The tests described above were all performed in sand. In soil very different conditions obtain. The compounds used may be destroyed by micro-organisms or other means, and may also be adsorbed by the soil particles. Several substances have already been tested in air-dried soil. Further sterilization made no apparent difference to the rate of disappearance of the compounds tested. Table 5 shows the activities obtained.

The effect of starvation on the threshold of orientation of 10 individual wireworms (A. sputator) which were known to be active was tested in the plate apparatus (in sand). A response to the boundary between water and the solution in question was taken as indicating activity of the latter (§ IV (iv)). For all 10 wirewonns when fed the activity of glucose was 2-1 but not 2-3, after 2 days’ starvation 2-3 but not 2-5, and after 4 days’ starvation 2-5 but not 2-7. After 7 days’ starvation the activity of glucose was 2-7 but not 2-9 for 9 wireworms, and 2-9 but not 3-1 for the remaining one. No further change occurred with 14 days’ starvation. The first track in glucose in Text-fig. 10 b is made by the last-mentioned wireworm, the other two tracks by wireworms for whom the activity of glucose was 2-7 after 7 days’ starvation. The orientation activity of glucose for wireworms thus increases with starvation parallel with biting activity if the latter be measured by the lowest concentration which causes any more biting than the controls (Text-fig. 7). For all 10 wireworms when fed the activity of asparagine was 9 but not 11. No change occurred with 2 days’ starvation, but after 4 days activity was 11 but not 13 for all. After 7 days activity remained at 11 for 7 wireworms but went to 13 but not 15 for 3, including the same wireworm for whom the activity of glucose went to 2-9. Of the tracks in asparagine shown in Text-fig. 10b, the first is in asparagine of concentration io^-11 and the last 2 that of concentration io^-13. No further changes occurred with 14 days’ starvation. The limits of activity reached may be taken as the limits of sensitivity of the receptors concerned for glucose and asparagine respectively. Sensitivity to glucose seems to have approximately the same value whether biting or orientation is being exhibited.

While sugars and other bitten substances cause both reactions, other active substances cause only orientation. The wireworms must therefore be able to distinguish between these two groups. The sensitivity of the receptors to orientation substances other than sugars and other bitten substances is comparable to that of the most sensitive olfactory receptors, while their sensitivity to the biting groups is similar to that found in organs of taste. Man is sometimes classed as ‘microsmatic ‘yet Bach (see Moncrieff, 1944, p. 80) records that for the human nose synthetic musk, vanillin and skatol have activities of the order of 10 or 11. Truly ‘macros-matic’ mammals do not appear to have been subjected to critical olfactometer tests and such experiments would not be easy, but the following are the activities of some of the most active substances stimulating the olfactory receptors in some other animals: ethyl butyrate for the newt (Triton cristatus) in air and water 5-4 (Gertz, 1938), bromostyrene for the antenna of the honey bee (Apis melliphora) 3-3 (von Frisch, 1919), and skatol for the water beetle (Hydrophihispiceus) in water 6-2 (Ritter, 1936). Bromostyrene has an activity of 4 for man and the newt, i.e. about equal to that for the bee (Gertz, 1938). It is probable that the latter is much more sensitive to other naturally occurring odours. The activities of sucrose for the organs of taste of these animals are as follows: for the human tongue 2-2 (Parker, 1922, p. 143), for the newt’s mouth 1-7 (Giersberg, 1926), for the honey bee’s mouthparts 1-7 (von Frisch, 1934), and for the palps of Hydrophilus piceus 2-5 (Bauer, 1938). The wireworm receptors which respond to orientation substances may be presumed to facilitate the discovery of ‘distant ‘sources of emanation, and it is interesting to note that these respond not to odours but, like those of aquatic animals, to dissolved substances at very low concentrations.

At first sight it may seem that orientation in substances which elicit biting may be caused merely by the wireworms stopping to bite, and that they elicit not two but only one response. But, as shown above, at certain times of year wireworms do not orientate to substances which normally elicit this response (Text-fig. 2). During such periods it was discovered that while wireworms would orient neither to sucrose nor to potato juice, they exhibited the full biting response to these substances. The average number of bites per 20 wireworms in 24 hr. was 582 (six experiments) on potato juice and 400 (three experiments) on 5 % sucrose. Individual wireworms have also been observed to give the orientating response at a boundary of glucose in the plate apparatus, while not biting glucose on filter papers. Biting and orientation must therefore be distinct, and the response exhibited depends upon the physiological state of the animal. Wireworms may at different times exhibit both responses, bite without orientating, orientate without biting, or exhibit neither respónse.

The mechanism of orientation of active wireworms to asparagine, glucose, sucrose and peptone was examined in the plate apparatus with soil as the medium. The wireworms had been starved for 2 days. To see whether the speed of movement was less when the sense organs were being stimulated by an active substance, the length of track made in 1 hr. by the same wireworm (A. obscurus-lineatus) was measured first in water then in the substance. Table 6 shows that the average lengths of tracks made in an hour in each of the substances tested is less than that in water. If it be assumed that track length is proportional to speed of movement, it follows that wireworms exhibit orthokinesis to each of these substances (Fraenkel & Gunn, 1940). Orthokinesis forms an integral part of the responses of wireworms to humidity (Lees, 1943 a) and temperature differences (Falconer, 1945 a).

In the choice chamber experiments a reduction of velocity on the bait side will lead to collection on that side. The situation is analogous to a vessel of gas divided by a non-conduction membrane with the separate masses of gas at different temperatures. For the pressure on each side of the membrane to be equal, corresponding to an equilibrium in the choice chamber in which an equal number of wireworms cross the boundary in opposite directions in unit time, the number of molecules in unit volume on the colder side must be greater. Without knowledge of the distribution of velocities in a wireworm population it is impossible to calculate the average retardation which will cause a given percentage to collect, but by making certain assumptions an approximation to the reduction in velocity can be found.

It is assumed that the velocity of all the wireworms on the water side is v and on the bait side ar, that the change in velocity occurs immediately the dividing line is crossed, and that in the equilibrium state there are a wireworms on the water side and b on the bait side.

So the degree of orthokinesis, w/v, required to give a collection of 64% under these simplified conditions is 0-56. This shows fair agreement with the degree of orthokinesis measured in plate experiments, but since the wireworms used in the latter experiments had been specially selected and the degree of variation in such experiments is great, the other reactions which have been shown to occur at the boundary are not excluded by this calculation from playing an important part in causing collection on the bait side of the choice chamber.

To examine the possibility of other orientating mechanisms a boundary between water and solution was arranged in the soil in the plate apparatus. Wireworms were placed pointing towards the boundary sometimes on one side and sometimes on the other. They were then left in the dark and when they had wandered about for a suitable time (usually 2 hr.) the tracks were traced. Some individuals were indifferent to the boundary. Others responded as described below, examples of tracks made by those which, having started on the water side and crossed to the bait side, turned again and responded at the boundary, being shown in Text-fig. 9. The movements of wireworms were also observed continuously in dim light. The usual behaviour was as follows. When proceeding from water into an active substance no response occurred at the boundary. But when proceeding from an active substance into water, the wireworms either backed along their tracks sometimes quite long distances and then sometimes made a side turning; or else proceeding forwards they executed a turn which brought them back into the active substance. Both forms of behaviour are to be seen in the tracks shown in Text-fig. 9. Sometimes wireworms moving out of the bait did not respond immediately at the boundary but went quite long distances into the water side before returning (cf. Text-fig. 10). The most reliable criterion for activity is therefore whether the wireworms are found on the bait side after two hours. This occurred every time with active wireworms and active substances. The responses to asparagine, glucose, sucrose, peptone and triolein all have the same character.

Lees (1943 a) and Falconer (1945 a) describe similar behaviour of wireworms at the boundary between two different humidities and temperatures, respectively. Both authors find difficulty in placing the behaviour exhibited in any of the categories differentiated by Fraenkel & Gunn (1940). Both argue against klinokinesis, for there is no evidence of increased turning on crossing the boundary from the preferred side, and Lees (1943 a) is further impressed by the supposition that turning would be impossible in nature since the animal is restricted by the limits of the burrow. He inclines to the view that the mechanism involved is klinotaxis, a directed movement involving the ‘comparison ‘in successive intervals of time of the intensity of the stimulus to which the animal is reacting. But the forward turn sometimes executed when the wireworm is crossing the boundary from water into an active substance (and there is evidence for the same movement in the tracks published by Lees (1943 a) and Falconer (1945 a)) may be regarded as an example of an increased rate of change of direction, and also involves the construction of a new burrow. On the other hand, the reversing movement along the burrow back from the boundary is probably best regarded as klinotaxis. The facts do not permit a decision in favour of either klinotaxis or klinokinesis, and it is perhaps wisest to refrain from forcing the behaviour observed into either category (cf. Gunn & Walshe, 1942). It must be remembered that these categories were devised for the description of the behaviour of animals in media which do not restrict lateral movement, and that they represent a limitation imposed by the animal’s own structure and sense organs and by the nature of the environmental situation.

Wireworms exhibiting orthokinesis responded to the boundary, and vice versa. Some wireworms were found to exhibit neither response. The responses of A. sputator and A. obscurus-lineatus are similar. In Text-fig. 9 the first track drawn for each substance is made bÿ a sputator and the others by an obscurus-lineatus individual. Responses in soil and sand are also similar (cf. Text-figs. 9, 10).

Wireworms respond to a gradient of concentration of both glucose and asparagine. A /boundary was arranged in sand in the plate apparatus between solutions of different concentrations. Active individuals (A. obscurus-lineatus) known to be responding to the weaker solution were placed in the stronger solution pointing towards the boundary. A response occurred with various differences in concentration. The tracks made when the gradient was between 1-26% (activity 1-9) and °’8% (activity 2-1) glucose solutions, and between asparagine solutions of concentrations 10⁹ and io^-3 respectively, are shown in Text-fig. 10 A.

Field experiments were carried out for two purposes: to discover (i) the effectiveness in the field of baits containing active substances, and (ii) the extent to which wireworms move about under natural conditions.

The baits consisted of blocks of paper pulp roughly 6×3×3 cm. soaked in solutions or emulsions of an active substance. Blocks soaked in water were used as controls. Slices of potato were also tested as baits. The baits were distributed at random, as recommended by Fisher & Yates (1943), in holes 10 cm. deep and 16 cm. apart. The numbers of each kind of bait and of controls were equal. They were dug up at intervals of 2, 3 or 4 days after burial and the wireworms on each counted and removed.

Preliminary experiments were conducted in a badly infested greenhouse on chalky medium-heavy loam at Fulboum. Six baits containing 2% glucose and 0-1 % asparagine caught 35 wireworms (30 A. obscurus-lineatus, five athous) and the six controls 1 athous in the first 2 days. After 4 days baits and controls had each caught i more obscurus-lineatus. Potato slices were then put down in place of the baits and after 2 days caught 11 obscurus-lineatus and 3 athous to the controls none. Wireworms were therefore still present in the soil after the bait had become ineffective, suggesting that it must have been destroyed.

A series of experiments were now carried out with baits of various substances in the ground at the same time. The experiments were performed during April and May 1945 in a market garden on heavy boulder clay near Hardwick which had a high population of wireworms. Glucose as well as the water controls was present in every combination of baits, so that all the other substances or mixtures of substances may be compared with this. The results are shown in Table 7. It is clear from the first three experiments that glucose and glucose + asparagine are equally effective and catch more than the controls, while asparagine catches no more than the controls.

In a further experiment it was shown that 5 % asparagine is also ineffective. Glucose and glucose + asparagine have ceased to catch more than the controls by about 8 days. Potato slices, on the other hand, continue to be effective beyond this period, and in fact remain so as long as they are alive (see below). They catch at a lower rate than glucose. Exp. 3 shows also that when wireworms are not removed they will continue to accumulate on potato baits. One difference between glucose and asparagine is that whereas the former elicits biting as well as orientation the latter elicits only orientation. The explanation of the ineffectiveness of asparagine may then be that although it causes orientation it provides no food, so wireworms will not remain in its presence. Exps. 4 and 5 show that triolein has about the same effectiveness as glucose but lasts longer. Peptone is ineffective in spite of causing both orientation and biting in the laboratory, and so is tannin. Potato juice is ineffective until it is twice concentrated. There are reasons for believing that the failure of these baits is due to soil action. Peptone and potato juice are both very favourable substrates for bacteria. This may explain why the threshold of glucose in potato juice on blocks in the field is four times the threshold of glucose alone. Tannin would be precipitated by calcium compounds present in the soil. The wireworms caught in Exp. 6 were tested in the laboratory and found to collect in o-1 % asparagine, in four trials 149 (= 63 %) out of 236 collecting on the bait side of the choice chamber.

Two further experiments were performed during June and July 1945 in a field on heavy gault clay near Barton. In the first the effectiveness of water, 2 % glucose, i % triolein and 1 % triolein + 0-5 % asparagine, were compared. In five trials with eight blocks of each kind representing forty baits of each kind and forty controls, the total numbers caught in 2 days were: control 30, glucose 99, triolein 147, triolein and asparagine mixture 169, representing 0-75, 2-5, 3-7 and 4’2 wireworms respectively per bait in 2 days. Triolein may be slightly more effective than glucose, but there is no significant difference between the last two figures, suggesting that asparagine does not enhance the effectiveness of triolein. In the second experiment different concentrations of glucose were compared. In 2 days the controls and 0-5, i, 2, 4 and 8% glucose baits caught 8, 14, 13, 21, 16 and 19 wireworms respectively per eight baits, showing that the concentration of glucose has no effect during this period.

A preliminary experiment was carried out during March 1945 in the Fulbourn greenhouse in which lettuce were growing and which contained a high proportion of wireworms. Groups of twelve potato slices were buried in square patches 4-5 sq.ft, in area in different parts of the greenhouse. The patches were classified according to whether they contained well-grown lettuce, poor lettuce or no lettuce. The poor quality and absence of lettuce was suspected as being due to damage by wireworms. The baits were dug up at intervals of 2 or 3 days and the wireworms counted and removed. Table 8 shows that the wireworms were not uniformly distributed but that density of population is correlated with density of lettuce, and that the athous were evenly spread amongst the other species (obscurus-lineatus). The 13 % in C is not significant, since it is based on only 16 wireworms. There was no significant difference between the numbers found in (197) and between (178) the rows.

As the population was reduced the average rate of catching per patch (twelve baits) fell in A patches from 7-5 during the first 2 days to 2-t during the last 2 days, in B patches from 5-7 to 0-3, and in C patches from 2 to o. After 16 days the soil in four A patches and in four adjacent unbaited ‘A’ patches of the same area was dug out to the depth of 1 ft. and sieved. Twenty-six wireworms or 6-5 per patch were found in the baited ‘A’ patches and 17 wireworms or 4-1 per patch in the unbaited ‘A’ patches. Both sets of patches were now baited with twelve potato slices as before. During the next 8 days the four previously baited A patches yielded 20 wireworms or 2-5 per patch for 2 days and the four previously unbaited ‘A’ patches 29 wireworms or 3 · 6 per patch for 2 days. These must have wandered into the patches either from the side or from deeper down in the soil. The next experiment was designed on a more elaborate scale to investigate this point.

Fifteen square patches 4-5 sq.ft in area each containing potato slices were arranged in the market garden near Hardwick. The experiments were performed during April and May 1945. All vegetation was removed from the experimental area. The patches were of three kinds : A, control ; B, with the bottom sealed off at a depth of i ft. by glass plates ; and C, with the sides sealed to a depth of 1 ft. by glass plates.

The plates were inserted 2 days after the beginning of the experiment. The wireworms on the baits were counted and removed at intervals of a few days as in the previous experiment. Most of the potato slices remained alive during the whole period of the experiment; those which did not were replaced by fresh slices. Table 9 and Text-fig. 11 shows the total number of wireworms caught per patch in 2 days in A, B and C patches at successive points of time. The rate of catching decreases with time in all three, but faster in B and C than in A. The fact that the rate of catching falls rapidly and then continues at a steady level suggests that wireworms are wandering into the patches from outside. The smooth curve in Text-fig. 11 is calculated for the A patches on the assumptions that one-third of the wireworms in a patch are caught every 2 days, independent of population density, that the original density is 135 in five patches, that 6 wireworms wander into the patches in 2 days when the difference in rates of catching inside and outside the patches is 45 in 2 days, and that the number wandering in at any time is proportional to the difference in rates of catching inside and out. If these assumptions are true then the original population gradually decreases and the proportion among those caught of individuals which have wandered in increases until all those caught are such wanderers. A stable state is thus reached in which the rate of catching equals the rate of wandering. The average rate of catching for the last three records is 5 per 2 days, which means that the basic population should be 10. If the above assumptions are true the rate of wandering will be 5, i.e. all those caught are wanderers. It can be calculated in the same way that in all 92 wireworms have wandered into the patch. The total number caught is 212, and as shown above to remain. If a third of those present are caught the original population was 135, since 45 were caught in the first 2 days. The original population calculated from the above considerations about wandering would be 212 +10—92 = 130, which is a fairly close agreement. The total number caught in B is 150, the original population calculated as for A 132, and the number remaining 7, therefore the number wandering in is 150 + 7—132 = 25. The total number caught in C is 173, the original population 150, the number remaining 3, therefore the number wandering in is 173 + 3—150 = 26. The glass partitions at sides and bottom thus both reduced the amount of wandering, from which it may be concluded that wandering occurs from all directions, and that there was a population of wireworms in the deeper layers of the soil which was being tapped during the course of the experiment.

An experiment was now performed during May and June 1945 on Mr W. Jackson’s land on Swaffham Fen to observe directly the wandering of wireworms in the field. 500 wireworms were placed in a small hole 6 in. deep in a field known to be free of other wireworms. This was a fine light black fen soil, tightly packed below the top inch, which is an ideal medium for movement: all vegetation had been removed from the surface. Concentric circles were described round this point every foot up to 7 ft. from the centre, and the whole marked off into four quadrants. Rows of potato slices were placed along the circumferences of the concentric circles in alternate quadrants. These slices were examined at intervals and when, after 11 days, the outer ones began to catch wireworms the soil in the unbaited quadrants was dug out circle by circle and sieved to a depth of 6 in. then in part to a further 6 in. In the central hole 80 wireworms were found in the first 6 in. and 15 in the next 6 in. The numbers found in the unbaited quadrants of successive circles were 14, 10, 8, 11, 10 and 14 respectively, making a total of 67. The first circle of one quadrant was dug out to 1 ft. in depth, yielding 2 wireworms. The number found in this quadrant to 6 in. depth was 6. This suggests that about 20 % the wireworms are below 6 in. Assuming that 20 % of the wireworms were missed in sieving the total number in the two unbaited quadrants would then be 67+ 13 (20% of 67 missed) + 16 (20% of 78 below 6 in.) = 96. The same number may be supposed to be present in the baited quadrants (actually 78 were caught on the baits so there may have been more in the baited quadrants). The number found in the centre was 95 plus say 10% missed making 105. The total number accounted for is therefore 295 out of 500. The shape of a graph of wireworm density (wireworms per sq.ft.) plotted against distance from the point of dispersal suggests that some of the wireworms unaccounted for are outside the 7 ft. zone, though some may have been destroyed by predators. This shows that wireworms can move over 7 ft. in 11 days, i.e. their average rate of migration in a given direction may be as much as 20 cm. per day.

Burrowing through the soil seems to be the usual method of movement of wireworms, but it is of interest to note that on several occasions when collecting in the field we found wireworms attacking the stems of wheat just above the surface of the soil. The position of the attacks showed that the wireworms had not penetrated up the interior of the stem from below, but had come to the surface and then begun the attack.

It has been shown that plants in general contain substances giving two types of response. The orientation response is, in some cases, evoked by very low concentrations of the substance in solution. The biting response requires in general higher concentrations. The two responses were originally separated by finding larvae in a condition in which they showed the biting response only. The relation between sensitivity of the larvae to substances causing orientation and biting respectively, is quantitatively similar to the relation between the sensitivity in smell and taste in ourselves.

By examining a variety of substances it was found that the active chemical compounds in both cases were either related to or to be found among the compounds known to be present in the food plants.

When baits are used it is found that only the substances causing the biting response are effective in catching the wireworms. Further, the addition of an orientating substance to baits containing triolein does not increase the number caught. When triolein, which is insoluble in water, is compared in the field in dry soil as a bait with a solution of carbohydrate in water there is no significant difference in the numbers caught. This suggests strongly that the random type of burrowing is the main factor in such experimental conditions. Orthokinesis and the biting response together might thus account for all the results with artificial baits. This, however, would not exclude responses involving turning (klinotaxis and klinokinesis) playing a part in the finding of root systems under natural conditions.

The conception which it is necessary to define is that of ‘random ‘burrowing. The first point that arises is that in field conditions the whole volume of the surface layer of the soil will not be equally explored by the wireworms. This is indicated as follows. The wireworms are considered to have a semi-permanent system of burrows. The growing plants themselves will, by the nature of the growth of their roots, produce a system of potential burrows or lines of least resistance. This is because of the sloughing off of the cortex of the roots some distance behind the growing point, and because plant rootlets are continuously being eliminated from the living plant. Further, plant roots will make use of the burrows left by the wireworms, and also of the potential burrows left by the dead roots of previous plants. Therefore there will be a definite connexion between the wandering through the soil of both the wireworms and the growing roots of plants.

It is readily seen how this can, independently of any further mechanism, increase the chance of larvae being found within the boundaries of the root system of the plants. It might partly explain the results of ‘good cultivation’ in diminishing the effect of the wireworm populations of the soil, and further show how vulnerable young plants may be in some cases where the roots actually pass into the semipermanent system of burrows.

This would go far to account for the fact that in dealing with the chemical responses we can find no evidence of any effect due to percolation taking place through an appreciable distance in the soil. The baits used in the experiments must be considered as intersecting a number of potential burrows or lines of least resistance. The plant-root system is, however, very different from the baits used.

It is just at this point that we could see a necessity for the two chemical responses. It is the general impression that the larvae rapidly find the main axis of the root system and the neighbourhood of the most actively growing part of the whole plant. The assumption is now made that, during the growth of the root system, as the root hairs are continuously eliminated some distance behind the growing root tips substances of the type of asparagine and aspartic acid will be set free into the soil. This will very slightly increase the target area of all the growing roots at the outside of the root system as a whole, and from the properties of the orientating response we could then infer that the presence of aspartic acid, etc., will tend to keep the larvae within the root system as a whole.

It must be emphasized that the wireworms in the soil have no reliable or precise general method of orientation. They can avoid extremes of temperature, and humidity below saturation, but they are in the dark and they are not sensitive to gravity. Thus one would expect chemical methods of orientation to be necessary. From the experiments described and by making use of the assumptions stated previously we can give a more detailed explanation of the method of food finding than was hitherto possible.

Finally, it may be concluded that the prospects of diverting an attack from the critical early stages of a crop are encouraging. The best time for this would presumably be before the wireworms have had the chance to establish burrow systems directly associated with the plants. The efficiency of any baiting method can, on the conclusions we have drawn from the work here described, readily be seen to depend greatly on the type and condition of the soil.-To bring the bait method to its fullest practical efficiency in wireworm control the aim should be to produce a bait substance resistant to bacterial attack. This should then be mixed with a non-repellent contact poison. On these lines the method might have a wide application.

Photographs of filter papers bitten by wireworms. Those on the diagonals contain carrot juice, the others water (see text).