Do travelling band solutions describe cohesive swarms? An investigation for migratory locusts

. We explore several hypotheses for the swarming behaviour in locusts, with a goal of understanding how swarm cohesion can be maintained by the huge population of insects (up to 10⁹ individuals) over long distances (up to thousands of miles) and long periods of time (over a week). The mathematical models that correspond to such hypotheses are generally partial differential equations that can be analysed for travelling wave solutions. The nature of a swarm (and the fact that it contains a finite number of individuals) mandates that we seek travelling band (pulse) solutions. However, most biologically reasonable models fail to produce such ideal behaviour unless unusual and unrealistic assumptions are made. The failure of such models, general difficulties encountered with similar models of other migratory phenomena, and possible approaches to alleviate these problems are described and discussed. Received: 25 April 1997 / Revised version: 14 September 1997     

Juvenile hormone-controlled vitellogenin synthesis in Locusta migratoria fat body: Cytological development

Cytological development in the fat body of adult female Locusta migratoria, related to vitellogenin synthesis, has been studied by light and electron microscopy. In the newly-emerged adult, the cells are filled with lipid droplets, which indent the nucleus, and with fields of glycogen, while ribosomes and endoplasmic reticulum are scarce. Correlated with the onset of vitellogenin synthesis, about day 8 of adult life, the nucleus enlarges, lipid droplets and glycogen decrease, and rough endoplasmic reticulum and Golgi complexes become the most abundant organelles. These changes reflect a conversion of the principal role of the fat body from nutrient storage to the synthesis and secretion of protein. They are prevented by allatectomy, and restored by subsequent treatment with the juvenile hormone analogue, ZR-515. Late in the first gonotrophic cycle, about day 20, dense bodies, vesicle-containing bodies and lysosomes are seen, indicating recycling of cellular materials. Five days after ZR-515 treatment, when protein synthesis has declined, the rough endoplasmic reticulum appears in arrays adjacent to lipid droplets, possibly awaiting reactivation. By the use of ferritin-labelled antivitellin immunoglobulin, vitellogenin has been localized intracellularly in the RER saccules and Golgi vesicles, and extracellularly in channels between the folded plasma membranes, showing sites of accumulation and secretion of this protein.     

Processing of mechanosensory information from gustatory receptors on a hind leg of the locust

Gustatory receptors (basiconic sensilla) on the legs of the desert locust, Schistocerca gregaria, are innervated by chemosensory afferents and by a mechanosensory afferent. We show, for the first time, that these mechanosensory afferents form an elaborate detector system with the following properties: 1) they have low threshold displacement angles that decrease with increasing stimulus frequency in the range 0.05–1 Hz, 2) they respond phasically to deflections of the receptor shaft and adapt rapidly to repetitive stimulation, 3) they encode the velocity of the stimulus in their spike frequency and have velocity thresholds lower than 1°/s, and 4) they are directionally sensitive, so that stimuli moving proximally towards the coxa elicit the greatest response.The mechanosensory afferents, but not the chemosensory afferents, make apparently monosynaptic connections with spiking local interneurones in a population with somata at the ventral midline of the metathoracic ganglion. They evoke excitatory synaptic potentials that can sum to produce spikes in the spiking local interneurones. Stimulation of the single mechanosensory afferent of a gustatory receptor can also give rise to long lasting depolarizations, or to bursts of excitatory postsynaptic potentials in the interneurones that can persist for several seconds after the afferent spikes. These interneurones are part of the local circuitry involved in the production of local movements of a leg. The mechanosensory afferents from gustatory receptors must, therefore, be considered as part of the complex array of exteroceptors that provide mechanosensory information to these local circuits for use in adjusting, or controlling locomotion.     

Flight-phase and visual-field related optomotor yaw responses in gregarious desert locusts during tethered flight

Tethered flying desert locusts, Schistocerca gregaria, generate yaw-torque in response to rotation of a radial grating located beneath them. By screening parts of the pattern, rotation of the unscreened grating turned out to induce a compensatory steering (by pattern motion within transversally oriented 90° wide sectors) as well as an upwind/downwind turning response (by pattern motion within the anterior ventral 90° wide sector). The strength and polarity of responses upon the unscreened grating results from a linear superposition of these two response components. The results are discussed with regard to a functional specialization of eye regions.In a typical experiment, 3 consecutive flight-phases, assumed to mirror start, long-range flight, and landing of a free-flying locust, were distinguished. They may result from a time dependent variation of the polarity and relative strength of upwind/downwind turning and compensatory steering responses. Starting and landing phases were under strong optomotor control and were dominated by the high-gain compensatory steering. In contrast, the phase of long-range flight was under weak optomotor control resulting from a low gain in both of the two response components. The biological significance of this variable strength of optomotor control on free flight orientation of swarming locusts is discussed.     

The structure of bimodal chemo-, thermo-, and hygroreceptive sensilla on the antenna of Locusta migratoria

Summary Coeloconic sensilla on the antenna of Locusta migratoria were investigated electrophysiologically, labeled and then examined under the scanning or transmission electron microscope. They can be categorized into two main morphological types: 1) sensilla with wall pores and two concentric cuticular walls (wp-dw sensilla), and 2) sensilla lacking wall pores (np-sensilla).Hygroreceptors were observed only in np-sensilla. Olfactory receptors were observed in wp-dw sensilla, but not in np-sensilla. Cold-sensitive units were found in both types. Wp-dw sensilla with a cold-sensitive unit also manifest a fatty acid olfactory receptor. This cold-sensitive unit appears to be less sensitive to drops in temperature than the cold receptor combined with hygroreceptors in np-sensilla. Recordings from both types revealed up to three or sometimes four units that could be distinguished on the basis of impulse amplitude.The number of np-sensilla on each antennal segment ranges from 0 to 6. These sensilla are pegs 4–5 m long with a terminal swelling. The dendritic outer segments of two sensory cells fit tightly into these pegs. The dendrite of the third sensory cell ends at its base. All three are up to 1 m thick, unbranched, and densely filled with microtubules. The cuticular wall and socket of the peg are characterized by (1) an electron-dense inner layer surrounding the dendrites and continuous with electron dense clusters at the base of the peg, and (2) clefts in the cuticle at the insertion of the peg.The present observations favor the hypothesis that insect sensilla combining the triad of moist air, dry air and cold receptors generally lack pores. Specifical, though still alternative roles in stimulus transduction can be suggested for substructures demonstrated in the present and previous electron microscopic investigations.Supported by the Deutsche Forschungsgemeinschaft (SFB 4:G1/D)     

Development of blood-cerebrospinal fluid barrier to horseradish peroxidase in the avian choroidal epithelium

Summary Four neurons in the brain of the migratory locust were immunohistologically identified with an anti-met-enkephalin antiserum. The perikarya of two of these cells are located in the center of each of the two groups of lateral protocerebral neurosecretory cells. The fibres coming from these perikarya terminate in numerous immunoreactive ramifications visible at the periphery of both tractus I to the corpora cardiaca, through which pass the neurosecretory products of the pars intercerebralis. The other two cell bodies are located at the bases of the two optic lobes; their fibres enter the posterior part of the protocerebrum and ramify around the root of the nervus corporis cardiaci II, another area through which neurosecretory products pass. The topographic distribution of these met-enkephalin arborizations suggests that these four neurons may act as neuromodulators of the acitivity of the major neurosecretory cells in the brain of this insect.     

Peaking of response in locust visual interneurones to objects of small angular subtense

One compound eye of an immobilised locust viewed a large screen on to which were projected discs of light for periods of 2 sec every 40 sec. The spike response was counted concurrently in the DCMD and the next largest axon in the contralateral nerve cord connective. The average score for 10 trials, after correction for background was plotted for a series of discs subtending a range of angles from 0.05 to 84°.It was found that the response of the DCMD peaked sharply and consistently at a subtense of 0.3°, and fell away to a low plateau or to zero over the range 2° to 84°. The response could exceed background down to subtenses as low as 0.05° (3′ of arc). The response of the next largest axon also showed an early peak, but it was inhibitory and resembled a mirror image of that of the DCMD, although it did not always coincide, ranging from 0.2 to 0.3°. The response, by contrast with the DCMD, rose to a high level at large subtenses, forming a flat peak.No explanation in optical terms could be found for this peaking at small subtenses, and a scheme is proposed by which a peak response could develop by the interaction of excitatory and inhibitory processes in the optic lobe.The peak value of 0.3° corresponds with the resolution limit for moving periodic patterns repeatedly demonstrated by Burtt and Catton (e.g. 1962, 1969). Such peaking behaviour would serve to lift the response curve of the whole visual system at high spatial frequencies, and thus extend the resolution limit.With stepwise reduction in intensity of a small luminous target there was a steep fall in the DCMD response, but a similar reduction for a large target had only a small effect. This could be explained by assuming that excitatory processes were prevalent for small targets, subtending about 0.3°, whereas for larger targets the excitatory and inhibitory processes came into balance over a wide range of intensities, thereby stabilising the response, at a low level.     

Accuracy of blood cytological screening techniques for the diagnosis of a possible hematopoietic neoplasm in the bivalve mollusc,Mya arenaria

The two in vivo bleeding techniques currently in use in our laboratory to diagnose a hematopoietic neoplasm in Mya arenaria are: (1) phase-contrast microscopy with fresh unstained hemocytes, and (2) bright-field microscopy with Giemsa-stained hemocytes. All in vivo diagnoses were checked by histopathological studies on tissues of the same mollusc. For both methods the correct diagnosis (true + or true ?) was made in 94 out of 100 clams examined. A gradation of tissue involvement was observed in the diseased clams and the accuracy of the in vivo diagnosis is related to the disease severity. There is a positive correlation between the degree of tissue involvement and the number of circulating neoplastic cells. For this reason the more extensive the neoplasm the better is the ability to diagnose the neoplasm by the in vivo bleeding techniques. Depending on the percentage of neoplastic cells present in the hemolymph, the neoplasm was graded from level 1 to 5, with 5 being the most severe. In general, at level 1, the accuracy of a single in vivo diagnosis varied from 66 to 71% and at level 2, the accuracy of diagnosis varied from 76 to 93%, while at all other levels the accuracy was 100%. The percentage of diseased clams detected by the in vivo bleeding technique was 89–91% and the percentage of nondiseased clams detected was 95%. These values can be further improved by combining the two tests and/or through multiple bleedings. Between the two types of in vivo tests, the Giemsa-stained hemocytes provided better precision of diagnosis than the fresh unstained cells, although the differences were slight.     

Formation of the receptor system in the hind limb of the locust embryo

Summary The formation of the peripheral nervous system in the metathoracic limb bud of a locust embryo was studied by antibody [anti-(horseradish peroxidase); (anti-HRP)] labelling. At 50% of embryogenesis, the major neural routes from the periphery to the CNS are established. There are two waves of receptor cell genesis, (a) At about 35%, the first precursors of internal receptors emerge from the epidermal cell layer and at about 58% apparently all internal receptors are formed. At this stage, the number and arrangement coincide with those of the first larval instar, (b) At the 55% stage, anti-HRP-immunoreactive cells appear, which can be associated with exteroceptors (hairs, campaniform sensilla); these reach their final number and cellular constitution at 65%–70% embryogenesis. The two waves are correlated to the formation of the first and second embryonic cuticle. The results thus indicate that the genesis of the receptor system and its connections to the CNS is completed at about 2/3 of embryonic development.     

Rheology and synergy of κ-carrageenan/locust bean gum/konjac glucomannan gels

The rheology and melting of mixed polysaccharide gels containing konjac glucomannan (KGM), locust bean gum (LBG) and κ-carrageenan (KC) were studied. Synergy-type peaks in the Young's modulus at optimal mixing ratios were found for both KC/LBG and KC/KGM binary gels at a fixed total polysaccharide content (1:5.5 for LBG:KC and 1:7 for KGM:KC). The Young's modulus peak for KC/KGM was higher than for KC/LBG gels. The same stoichiometric mixing ratios were found when either LBG or KGM was added to KC at a fixed KC concentration, where the Young's modulus increased up to additions at the stoichiometric ratio, but leveled off at higher LBG or KGM additions. Addition of KGM or LBG to the 2-component gels beyond the stoichiometric (optimal) mixing ratio at a fixed total polysaccharide content led to a decrease in the Young's modulus and an increase in the rupture strain and stress in extension, and both trends were stronger for KGM than for LBG.