Introduction

Collembolan mouth-parts are described as entognathous, and it is generally recognised that there are two distinct types. In what is considered to be the more primitive or generalized type (Salmon 1964; Massoud, 1967) the mouth is a wide opening at the distal border of the head. The mandible is robust and carries an apical toothed area or "pars incisiva" and a broader molar plate provided with a number of rows of teeth, the "pars molaris" of Massoud (1967). It is assumed that forms with this type of mandible chew their food, although some doubt has been cast upon this interpretation by Goto (1972). In the second group the mouth region is tapered to form a cone shaped structure known as the buccal cone. The mandible lacks a molar or grinding plate and is described by Massoud (1967) in the Neanuridae, as reduced to a long, slender, flattened plate carrying an apical toothed area the "pars incisiva". This plate is said to have a cutting action although this does not seem to be in agreement with the theory accepted throughout entomological literature, that forms provided with a buccal cone suck their food. It seems possible however, that a suction method of feeding has been attributed to these forms because of the apparently sharply pointed nature of the cone, rather than on a knowledge of the fine structure of the mouth-parts.

Recent writers including Goto (1972) and Massoud and Ellis (1974) describe the fine structure of several chewing type Collembola, but the structure and function of the mouth-parts of forms lacking a molar plate is little understood.

Wolter (1963) while accepting the view that forms with a buccal cone suck their food, drew attention to differences in the structure of the maxillary heads. He considered that those with a well developed toothed maxillary head probably rasped their food while those with a stylet-like head probably pierced it, before sucking the fluid material through the cone.

The diet of forms lacking a molar plate is unknown. It was noted however, in feeding studies by Macnamara (1924) and Poole (1959), that in subsequent preparations of the gut of these particular species, there were rarely any visible gut contents. This seemed to suggest that page 2they are fluid feeders. Singh (1969) also observed mainly fluid gut contents in Friesea mirabilis (Tullberg) agreeing with the results of Poole (1959) on the same species. Particles identified in the gut included fungal spores and hyphae, yeast, bacteria and collembolan eggs.

Adams and Salmon (1972) in a study of Brachystomella parvula (Schaeffer), a collembolan lacking a molar plate, observed that a sucking tube is not formed during feeding and the mouth-parts could not be described as adapted for sucking. It was noted that the gut contents were mainly of a fluid nature containing structures resembling bacteria, encysted protoza and fine unidentifiable particles.

Recently attention has been drawn to the enigma in detritivore communities of a high degree of species diversity, accompanied by an apparently low degree of feeding specificity (Anderson, 1975; Joosse, 1975) and Collembola would seem to be no exception to this (Anderson and Healey, 1972; McMillan, 1975). It is suggested therefore that a greater knowledge of the fine structure and function of collembolan mouth-parts could provide more specific information on their feeding habits.

In this study the fine structure of the mouth-parts of Ceratrimeria leleupi, a species lacking a mandibular molar plate and formerly regarded as a piercing and sucking type, is examined using the scanning electron microscope.

Specimens are also examined under the light microscope for evidence of food in the gut.

Materials

A collection of Collembola from Zaire contained several hundred specimens of Ceratrimeria leleupi Salmon and Adams, 1979. In the majority of specimens examined, the buccal cone was closed and dissection of the cone and mouth-parts was necessary for identification purposes. In a few specimens however, the buccal cone was open and the mandibles and maxillae were exserted from the mouth. These specimens were selected for detailed examination of the mouth and mouth-parts in situ.

In order to view the mouth-parts in their natural extended position no attempt was made to dissect them out of the cone. Whole specimens were mounted with the buccal cone uppermost. The following description of the mouth-parts of C. leleupi is from observations with the scanning electron microscope.

Specimens were prepared using the critical point drying method. They were fixed to a standard Cambridge S.E.M. stub with silver dag, sputter coated with gold and viewed on a Cambridge Mk. 11.A scanning electron microscope.

Morphology Of The Mouth-Parts

The buccal cone projects antero-ventrally from the head and is formed from the distal borders of the labrum and labium united basally on the cone by the oral folds (Folsom, 1900). (Figs. 1, 2). The lateral free margins of the labrum are folded and thickened with smooth cuticle and converge towards the distal border where they end abruptly in an upturned collar-like structure (Fig. 1). The general tuberculate cuticle of the body is continued on to the base of the labrum. A series of longitudinally arranged folds of finely granular cuticle run forward on the labrum merging to a single fold at the distal border (Fig. 1). The page 3

Fig. 1 Ceratrimeria leleupi

Dorsal view of open buccal cone, showing mandibles protruding beyond edge of labrum and divided labium to right and left. The rod-like maxillae show partially at upper left, and the superlinguae appear at the distal border of the labrum. Note the folded, thickened, non-tuberculate edge of the labrum. Lb, labrum; Lm, labium; M, mandible; Mx, maxilla, (× 1,200).

page 4

Fig. 2 Ceratrimeria leleupi

Ventral view of open buccal cone. Superlinguae show beneath labium in upper centre of cone and a rod-like maxilla emerges from right of cone. Mandibles are partially obscured distal to superlinguae. Note deeply cleft labium with thickened folded borders, antenna above and at lower right. Sl, superlinguae; Lm, labium; Mx, maxilla; (× 600).

page 5labium tapers distally on the cone in a similar manner to the labrum and is divided medially into right and left halves (Fig. 2). The lateral and terminal borders, and distal portion of the medial borders are folded and thickened with smooth cuticle. From the medial division on the cone, the ventral line continues back the entire length of the labium to the thorax. The cuticle of the labium is tuberculate but in the distal half of the cone it is finely granulate and thrown into a number of irregular folds. Each half of the labial section of the cone carries a small basal seta and nine larger setae distal to this, one of which, situated near the basal seta is distinctly longer than the rest. The mouth opening at the apex of the cone extends down each side of the cone bordered by the thickened and folded margins of the labrum and labium. The mandible is a slender, elongated, rod-like structure with an expanded apical region which resembles a claw or scoop (Figs. 3, 4). Six strongly curved teeth of varying size are arranged at the apex and along the medial border. They consist of four major teeth with a small tooth between the first and second and another between the second and third major teeth. The basal or sixth tooth, is the longest and stoutest of all. The lateral border of the mandibular head is thickened and together with the apical tooth, is rolled slightly inward. The toothed border of the mandible is also rolled over in such a manner that an elongated cupped depression is formed between them, and the whole mandible head resembles a claw. The mandibular head is approximately 80 in length and 20 μm in its widest part. Each maxilla consists of a pair of slender rods or lamellae. In all specimens examined the rods were closely opposed (Fig. 3), one of the pair being slightly longer than the other and faintly hooked apically. The maxilla is approximately 10 μm in diameter distally. The maxillae when exserted usually extend slightly beyond the mandibles. No opening could be found at the distal extremity of the maxilla to suggest a chemoreceptor function. The superlinguae are a pair of large, lamellate, folded structures lying between the labrum and labium. They appear to be situated ventrally to the mandibles and maxillae and may be partly exserted from the mouth opening (Fig. 1).

Discussion

It is of interest to consider the function of these mouth-parts and what type of feeding is possible in this collembolan. The labium is divided into separate halves for almost the entire length of the cone and this division, together with the lateral extensions of the mouth allows the distal portion of the cone to open when the mandibles and maxillae may be extended beyond the mouth opening (Figs. 1, 2). As the maxilla is hooked apically it would not appear to act as a piercing stylet. The mandibular shaft is rod-like, roughly circular in cross section rather than a flat shaft and the head is not a flat cutting plate as previously described (Massoud 1967), but is modified to form a cupped, claw shaped structure. The mandibular head therefore cannot be termed a "pars incisiva" as the teeth do not form a cutting edge, and the action of the mandibles is no longer a transverse cutting action, but a protraction and retraction (Wolter 1963).

As previously discussed, Collembola of this type have long been regarded as sucking, or piercing and sucking their food. It was considered that owing to the small size of the apical mouth opening they are not capable of conveying any solid food particles into the mouth cavity. The consistent absence of visible gut contents observed by a number of writers, including Macnamara (1924) and Poole (1959), was page 6

Fig. 3 Ceratrimeria leleupi

Distal portion of cone with mandible and maxillae partly extruded. Lower maxilla shows double structure and mandibles carry bacteria. M, mandible; Mx, maxilla; Lb, labrum (× 2,500).

page 7

Fig. 4 Ceratrimeria lelupi

Mandibular head showing claw-like form and containing bacteria (× 5,000).

page 8considered to indicate that they are fluid feeders. The entire mouth cone was said to function as a suction tube attaching itself firmly to the surface of the substrate during feeding.

The mandibles and maxillae of the present species are clearly not sucking structures and with the loss of the molar plate comminution of food particles by the mandibles is not possible. The structure of the mouth opening with thickened folded borders and divided labium bears a strong resemblance to that observed in Brachystomella parvula (Schaeffer) by Adams and Salmon (1972). In B. parvula the terminal portion of the cone remains open during feeding and the maxilla heads and superlinguae make rapid movements in and out of the mouth openings. Specimens of the present species were found with the cone open and the mouth-parts extended to varying degrees. As noted above, in forms possessing a buccal cone, the action of the mandibular and maxillary muscles is protraction and retraction. It seems reasonable to

Fig. 5 Ceratrimeria leleupi

Distal end of maxilla, carrying bacteria of branching and coccoid form (× 7,000).

page 9assume therefore, that the action during feeding in this species consists, as in B. parvula, of an opening of the distal portion of the cone which allows the heads of the mandibles and maxillae to emerge and operate in a protraction-retraction movement conveying food particles in to the mouth. As the mandibles and maxillae are of minute size and lack grinding or chewing plate, the size of food particles or organism ingested is accordingly limited. It has been suggested by Christiansen (1964) that bacteria may form an important part of the collembolan diet and several species were cultured on bacteria as the sole source of food. These however were not visible within the gut after ingestion. In the present study, bacteria of bacillus and coccoid form were observed on the mandibular heads of a number of specimens (Figs. 3, 4). Both mandibles were found to contain bacteria with very little other material present. In other specimens, although the mouth-parts were often fully extruded, the cupped mandibular heads appeared empty. Bacteria were the only identifiable material found on the mandibles. Similarly, bacteria of coccoid and branched form were found adhering to the maxillae, concentrated particularly at the distal end (Fig. 5). Bacteria were not observed on any other parts of the body. The consistent lack of visible gut contents in this species and the presence of bacteria on the mandibles and maxillae suggests that small organisms such as bacteria are at least a source of food.

The question then arises as to the function of the buccal cone. If it does not act as a suction tube, does it serve some other function in food selection? It has been noted that in Collembola possessing a buccal cone, the structure and function of the mouth-parts has changed. The molar plate with its grinding action is lacking and the mandibles, if present, are modified to slender, elongated, extremely delicate structures. With this modification there is a consequent reduction in the size of food particles ingested. The narrowing of the wide frontal mouth region to a tapered buccal cone appears to have evolved with this structural change in the mouth-parts and it seems probable that the cone serves to support and protect the delicate protrusible mouth-parts.

Summary

The mandible is seen to be a cupped claw-like structure. It is not a piercing stylet, as might be expected in an insect regarded as piercing and sucking its food, nor is the head in the form of a flat cutting plate. No locking structures are found on the cone which might serve to hold the labrum and divided labium together to form a suction tube during feeding. The maxillae, formerly described as stylet-like are seen to be hooked apically and would not appear to be adapted for piercing. They are finely delicate structures, but it is possible they may serve to supplement the function of the dentate claw-like mandibles in gathering food. The size of the apical portion of the maxillae, 1-2 μm in diameter, is comparable to the size of a generalized soil bacterium 1 μm³ (Clark 1967). As the maxillae extend beyond the mandibles in what appears to be the feeding position, they would seem suited to rake or sweep small food particles such as bacteria in to the cupped mandibular head. These food particles could then be conveyed to the digestive tract by the protractor retractor action of the mandibles.

The present study indicates that this collembolan does not feed by sucking or chewing its food but it could be described as browsing on the microfauna or flora of detrital material. This supports the view of Adams and Salmon (1972) that the division of Collembola into chewing forms and sucking forms is no longer valid.