Grooming Behavior in Diplura(Insecta: Apterygota)

Insect grooming studies are adding an important new dimension to knowledge of comparative behavior and evolution. Recent advances include an overview of a few selected movements of insects and myriopods (Jander, 1966), studies of the functional morphology of grooming structures (Hlavac, 1975), extensive reports about individual orders (Coleoptera: Valentine, 1973; Hymenoptera: Farish, 1972), quantitative studies at species levels (Chironomidae: Stoffer, in preparation; Drosophila: Lipps, 1973), and many less inclusive works. All such studies have difficulties which include the inability to know when an observed sequence is complete, the enormous number of potential taxa, the problem of generalizing about families and orders from small samples of individuals or species, and the absence of data from primitive or odd groups which may be critical for interpreting evolutionary sequences. The first three difficulties can be partially solved by increasing sample sizes and combining observations; however, the fourth can be solved only by availability. Grooming in the apterygote order Diplura is a good example because we can find only incomplete reports on one species. Recently, we have studied ten live specimens representing two families and three species; the data obtained provide an important picture of grooming behavior in one of the most primitive surviving orders of insects. Our observations greatly extend the limited discussion of grooming in the European japygid Dipljapyx humberti (Grassi, 1886) reported by Pages (1951, 1967). Data on Dipljapyx are incorporated here, but have not been verified by us. Initially we asked two questions: The first concerned whether a very primitive insect would enable us to observe a primitive grooming repertory; what we actually observed were primitive insects with grooming behavior beautifully tuned to a special and restricted environment. The second question concerned the effects

Insect grooming studies are adding an important new dimension to knowledge of comparative behavior and evolution. Recent advances include an overview of a few selected movements of insects and myriopods (Jander, 1966), studies of the functional morphology of grooming structures (Hlavac, 1975), extensive reports about individual orders (Coleoptera: Valentine, 1973; Hymenoptera: Farish, 1972), quantitative studies at species levels (Chironomidae: Stoffer, in preparation; Drosophila: Lipps, 1973), and many less inclusive works. All such studies have difficulties which include the inability to know when an observed sequence is complete, the enormous number of potential taxa, the problem of generalizing about families and orders from small samples of individuals or species, and the absence of data from primitive or odd groups which may be critical for interpreting evolutionary sequences. The first three difficulties can be partially solved by increasing sample sizes and combining observations; however, the fourth can be solved only by availability. Grooming in the apterygote order Diplura is a good example because we can find only incomplete reports on one species. Recently, we have studied ten live specimens representing two families and three species; the data obtained provide an important picture of grooming behavior in one of the most primitive surviving orders of insects. Our observations greatly extend the limited discussion of grooming in the European japygid Dipljapyx humberti (Grassi, 1886) reported by Pages (1951,1967). Data on Dipljapyx are incorporated here, but have not been verified by us.
Initially we asked two questions: The first concerned whether a very primitive insect would enable us to observe a primitive grooming repertory; what we actually observed were primitive insects with grooming behavior beautifully tuned to a special and restricted environment. The second question concerned the effects *Manuscript received by the editor September 26, 1978.

Psyche
[June-Septenber of endognathous mouthparts on grooming. The invaginated, noncondylar mandibles and maxillae of Diplura might reduce their effectiveness in oral cleaning, and result in an increased importance of leg rubbing movements. In fact, leg rubbing was seldom observed.
The rarity of rubbing has two possible explanations: either endognathy does not significantly modify grooming or else most leg rubbing movements have not yet evolved in Diplura.

MATERIAL EXAMINED
Campodeidae (seven specimens and seven hours of recorded observations plus about five additional hours of non recorded observation which add no new data) Ohio, Franklin Co., Columbus, Upper Arlington, 20 September, 1975, B. D. Valentine family, in soil in back yard (1 specimen). Same data except 5 November, 1977, in soil under boards and logs in back yard (6 specimens). Many additional specimens were seen and collected by breaking up clods of dirt in a garden. Japygidae (three specimens and nine hours of recorded observation plus about four more hours which duplicate previous data). Alabama (1957). Silvestri (1933a) and Paclt (1957) list two species of this subgenus occurring east of the Mississippi River, Campodea (C.) fragilis Meinert, 1865, and Campodea (C.) plusiochaeta Silvestri, 1912. Both are illustrated and described by Silvestri (1912). Our specimens more closely match C. plusiochaeta because the cercal setae are fairly long on all segtnents, as opposed to the long basal and shorter distal cercal setae of C. fragilis, and because there are bifurcate antennal setae, as opposed to the serrate or plurnose setae of C. fragilis. Nevertheless, the determination is not firm and the specimens should be listed as Campodea (Campodea) ? plusiochaeta Silvestri, 1912. The Ohio japygids key in Paclt (1957) to the genus Metajapyx Silvestri, 1933. Using Snith andBolton (1964)

RESULTS
CLEANING. Involves grootning with the mouthparts.
Antenna Clean. Passage of the antenna through the mouth is accomplished in two major modes: unassisted and assisted. In unassisted, which is the usual mode in Diplura, the antenna deflects into the tnouth due to its intrinsic tnusculature, and the legs are not involved. In Campodea this novenent is vertical to the substrate, the antenna is curled ventrally under the head and is chewed by the mouthparts; in japygids the rnovetnent is rarely vertical, the antenna usually is curled along a tnore horizontal plane from an initial position lateral of the head, and is usually drawn rapidly through the open mouthparts; less frequently it is chewed by the tnaxillae. In the nuch rarer assisted tnode, the ipsilateral foreleg pulls the antenna into the mouth and in both families is either returned to the substrate or held in tnid-air; in addition, the japygids were occasionally observed using the ipsilateral foreleg to help hold the antenna in the mouth by placing the leg crosswise in front of the mouthparts.
Pag6s (1967) points out that in Dipljapyx the foreleg holds the antenna during chewing by the naxillae, but is not used when the antenna is drawn through the maxillae without chewing movements. Palp Clean. A maxillary palp is passed unassisted through the rnouthparts in the anterior mode in which the palp tip projects posteriad and is drawn anteriad out of the mouth. This was observed clearly in Metajapyx. (In Campodea, maxillary palpi are one segmented and the labial palpi are vestigial.) Foreleg Clean. A foreleg is raised and extended forward while the head turns to the side to reach it; the leg is essentially in a ventrolateral position during cleaning, and is drawn posteriorly through Psyche [June-September the mouth, tarsal claws last. This occurs in both families and all three genera.
Midleg Clean. A nidleg is brought forward alongside the body and the head turns and dips to reach it, the linb noving posteriorly through the nouth, tarsal claws last. There are three modes: under L1, in which the foreleg is raised out of the way, in both fanilies; L1 pull, in which the raised foreleg is used to pull the nidleg into the nouth, seen rarely in Campodea and reported in Dipljapyx by Pag6s (1967); and over L1, where the foreleg renains on the substrate and the rnidleg crosses above is, seen in Metajapyx.
Hindleg Clean. A hindleg is brought forward alongside the laterally arched body and the head turns and dips to reach it, the limb moving posteriorly through the mouth, tarsal claws last. There are three nodes in Diplura: under L+2, where fore and mid legs are raised out of the way, in both fanilies; under L, over L2, which is self-explanatory and occurs in both fanilies (in this node both fanilies usually raise and partly extend L1, and japygids sornetines flex L and position it under the body); and L pull in which the foreleg helps pull the hindleg to the nouth, in Campodea and Dipljapyx.
Fore-Midleg Clean. Ipsilateral fore and midlegs are passed simultaneously through the mouth in anterior-posterior sequence. This infrequent action occurs in both Campodea and Metajapyx. Sometimes both tarsi are involved, but usually the fore tarsus and mid tibia are the parts cleaned. Fore-Hindleg Clean. As above, the ipsilateral limbs moving posteriorly through the mouthparts, observed rarely in Campodea.
Mid-Hindleg Clean. As above, except that the movement seems to be a rare continuation of Hindleg Clean, under LI+2, where the midleg becomes involved; in no case was the movement initiated independently of Hindleg Clean. This movement was observed rarely in Metajapyx.
Body Clean. Both families can bend double and use their mouthparts to groom body surfaces from the thorax to the cerci. These movements are less frequent than other grooming, so it is not known if the differences between the two families are real or sampling error. Watching these animals, the observer rapidly gets the impression that they can probably reach any body part they wish except the pronotum. At present, the campodeids have been seen cleaning all three coxae with the head directed ventro-posteriad; they also clean the lateral edge of the body, the styli, and the cerci with the body curled laterally. Cercal grooming techniques appear to be very diverse and are more controlled by position and substrate irregularities than by a stereotyped program. For example, the cerci can be held by L1, or by L1+1, or by L1+2, in each case the remaining ipsilateral legs are under the cercus; other variants involve L3 raised out of the way, L2 raised out of the way, and the cercus positioned over all three ipsilateral legs. Body cleaning in japygids extends at least from the mesonotum or mesosternum to the cerci, including dorsal, lateral, and ventral surfaces; during cercal grooming, the mouth can work the outer margin of a forceps from base to apex, around the tip, then the inner margin to and across the anal area, and out the inner margin of the contralateral forceps to its tip; the far outer margin is not groomed until the insect straightens and bends to the opposite side. In Dipljapyx Pag6s (1967) reports that the thoracic legs hold the abdomen when the body is tightly curved to clean from the mesothorax to the fifth abdominal segment.
RUBBING. Involves progressive contact of body parts with each other or with the substrate. In Diplura, all rubbing is of low frequency.
Antenna-Foreleg Rub. The fore tarsus or tibia is used to rub the dorsal surface of the ipsilateral antenna. This occurs in Campodea where the movement is confined to the basal antennal segments, and is sometimes combined with and precedes Antenna Clean, assisted.
Head-Foreleg Rub. In Campodea, the fore tarsus is used to rub the venter of the head and the mouthparts; in Dipljapyx, Pag6s describes head capsule rubs but does not indicate the areas involved.
HeadoMidleg Rub. Also in Campodea, a midleg is used to rub the renter of the head.
Head-Substrate Rub. Dipljapyx was observed rubbing the labial region of the head on the substrate with a sideways motion.
Body-Midleg Rub. In Metajapyx, the midleg is used to rub the dorsal and lateral surfaces of the thorax.
Body-Midleg-Midleg Rub. Also in Metajapyx, this is the bilateral version of the previous movement, both midlegs rubbing different thoracic regions simultaneously.
Body-Hindleg Rub. In Metajapyx, the hindleg is occasionally used to rub the dorsal or lateral surfaces of the thorax. Body Rubs can be combined, for on one occasion the thorax was rubbed Psyche [June-Septetnber simultaneously by a mid and hind leg from opposite sides.
Body-Substrate Rub. Pages (1967) reports that Dipljapyx rubs the thoracic sternum and abdominal base energetically on the substrate. He recognizes that this may be territorial marking, but believes that grooming is more probable.

DISCUSSION
The grooming patterns of Carnpodeidae and Japygidae are basically sinilar with one najor exception. In Antenna Clean, canpodeids chew the antenna with the naxillae during passage through the mouth, while japygids usually open the nandibles and then scrape the antenna rapidly through the open naxillae without chewing notions. Japygids can also chew the antenna but do so less frequently. The differences in grooning suggest different tnaxillary structures. Dissection of Metajapyx reveals an extraordinarily complex lacinia with five pectinate lanellae along the mesal face. Since we believe that these lanellae are the structures with which the antennae are cleaned; their distribution and function should be considered in future dipluran studies. Illustrations of these structures can be seen in the following works.
Japygidae: (note that the last genus is sonetirnes listed in a separate fanily).
Jander  states that grooning "... the antennae and all of the legs with the mouthparts.., is... to be regarded as the prinordial node of grooming..." in tracheate arthropods. It is true that oral cleaning novenents predominate in diversity and frequency in prinitive taxa, but it is also true that virtually all prinitive arthropods have rubbing novernents too. In nost cases it is inpossible to decide objectively which cane first.
Many factors affect grooning, and all act on both primitive and derivative taxa. For example, grooning novenents have constraints imposed by body flexibility and degree of leg novement. The configuration of a coxa and its cavity can be prinitive or derivative, but superimposed on this basic structure are the results of selection for plane of leg novenent, rotation, strength, speed, body height, and grooning requirements. The resolution of these diverse pressures nust result in a norphological cornpronise which affects grooning capability, but has little to do with prinitiveness.
Additional exanples are numerous. An elongate, flexible, softbodied organisn has different grooning patterns frown a fatter, nore rigid, sclerotized organism; one with easily abraded scales will Psyche [June-Septetnber be different frotn one with firtn setae; and an interstitial inhabitant will be different frotn a subcortical or leaf-litter inhabitant. The point is that all of these kinds of organistns occur in Apterygota and all are atnong the tnost pritnitive known hexapods.
In Diplura, grootning of the antennae, tnidand hindlegs involves at least fourteen cleaning positions, all of which appear to be satisfactory. This diversity is quite retnarkable and is unequaled in other insects (Valentine, unpubl.). The grootning of dipluran forelegs involves only one tnode. The stereotypy of foreleg grootning contrasts sharply with the diversity of antennal, tnid-and hindleg grootning. The logical explanation is that the single foreleg technique works in tnost or all situations, while no one technique works for the other appendages. Enviromnental constraints appear to require that the insect reach and grootn its antennae, tnid-and hindlegs in several alternate ways. Diplura are basically interstitial organistns. Altnost all specitnens were found in the soil under undisturbed stones or boards, or in soil clods in gardens. A standard technique for finding catnpodeids was to break up the datnp clods in a freshly plowed field or while digging potatoes. The very fine tunnels and cracks in this unyielding substrate are inhabited principally by stnall tnyriapods, Colletnbola, and Diplura. Since catnpodeids do not burrow and japygids do so very weakly (Pag6s, 1967), they pritnarily use the interstices already present. In such a habitat body configurations are subject to an infinite diversity of living spaces. A grootning behavior possible in one crack tnay be itnpossible in another; however, a tnodification tnay work. We believe that the unequal grootning diversity in Diplura is a response to the probletns of an interstitial life style. Foreleg grootning, where the leg is sitnply raised to the tnouth, does not require any special bending or tnovetnent, so one technique does the job. Antennal, tnidand hindleg grootning require unusual movetnents of the appendage or of the body. Such tnovetnents tnay be litnited by the varied configurations of the crawl space, and tnust accotnmodate to those configurations; thus, a variety of alternate positions appears to be a necessity.
It is itnportant to contrast the retnarkable freedotn of grootning positions of Diplura, with the very high degree of stereotypy in such orders as Diptera and Hytnenoptera. The point is that a discussion of insect grootning based on Diptera or Hytnenoptera is as biased towards stereotypy as a discussion of Diplura is biased towards lack of stereotypy. Present literature etnphasizes the stereotyped aspects of grooning, but it should be obvious that generalizations based on highly derivative or pritnitive orders are not valid for the entire class and nay be skewed in opposite directions. The order Thysanura would be a heuristic study because of the diversity of surface textures. There are scaly lepistnatids, catnpodeid-like nicoletiids, and sclerotized, non-scaly lepidotrichids. Grooning in these three fanilies tnay further clarify why the degree of stereotypy varies frotn taxon to taxon. LITERATURE CITED FARISH, D. J.