Review of larval morphology of beetles of the suborder Archostemata ( Insecta : Coleoptera ) , including first-instar chaetotaxy

This paper presents a synthesis of morphological information on larvae of the beetle suborder Archostemata. Larvae of the following families and species were studied: Ommatidae: Omma sp.; Micromalthidae: Micromalthus debilis LeConte, 1878; Cupedidae: Priacma serrata LeConte, 1861, Distocupes varians (Lea, 1902), Rhipsideigma raffrayi (Fairmaire, 1884), Tenomerga cinerea (Say, 1831) and Tenomerga mucida (Chevrolat, 1829). Morphological characters of the suborder and three families are described. Monophyly of the suborder is strongly supported by more than 10 larval autapomorphies. A close relationship between Micromalthidae and Cupedidae is confirmed. New larval characters are introduced, including chaetotaxy of first instar larvae of Micromalthus LeConte, 1878, Priacma LeConte, 1874 and Distocupes Neboiss, 1984. An identification key to families and subfamilies of Archostematan larvae is provided, along with a checklist of extant Archostemata taxa. The work is illustrated with 120 morphological drawings.

archostematan taxa is provided.Phylogenetic affinities based on larval morphology of some archostematan taxa discussed, however no formal cladistic analysis was performed.Some peculiar morphological characters are discussed in detail and new larval synapomorphies of the suborder suggested.

MATERIAL AND METHODS
This study is based on examination of Archostemata larvae kept in the following collections (names of curators in parentheses): ANIC -Australian National Insect Collection, Canberra, Australia (J.F.Lawrence, S.A. lipi  Larvae were disarticulated, macerated in hot KOH and mounted on microscope slides either temporarily in glycerine, or permanently in Euparal.A compound microscope, MBI-2, with up to X900 magnification was used to study the larval morphology.Drawings were made with the aid of a camera lucida.Morphological terms used in this work are those explained by Lawrence (1991: 147-177).The concept of the order and families follows Lawrence & Newton (1982, 1995).The use of the terms "seta" and "pore" follows that of Bousquet & Goulet (1984) for Carabidae; Alarie & Balke (1999) for Dytiscidae; Ashe & Watrous (1984) and Thayer (2000) for Staphylinidae; Kovarik & Passoa (1993) for Histeridae; Grebennikov & Beutel (2002) for Ptiliidae; Delgado & Soler (1996, 1997) for Hydraenidae and Wheeler (1990) and Kilian (1998) for Leiodidae.

Description
First-instar larvae: Like older instars, except general appearance is more of the "campodeiform" type.Head fully protracted.Frontal sclerite separated from epicranial plates by clear frontal suture (Priacma) or frontal sclerite fully united with epicranial plates and frontal suture not detectable (Micromalthus, Distocupes).First instar larvae of Rhipsideigma and Omma are unknown.Lateral sides of cranium with or without single stemma.Egg-bursters absent.Frontoclypeal suture not detectable.Body segments similar in shape, with (Distocupes) or without (Micromalthus, Priacma) transverse membranous folds dorsally and ventrally.Chaetotaxy (most full set of sensilla is indicated, reductions are common and noted for each genus separately): Cranium with setae 1-24 and pores A-R; prothorax with setae 1-21 and pores A-B; mesothorax with setae 1-23 and pores A-B; abdominal segments I-VIII with setae 1-16 and pores A-B; abdominal segment IX with setae 1-17 and pores A-C; abdominal segment X with setae 1-4.Older-instar lar-vae: Body cerambycoid (less so in Omma), with tergal ampullae on thorax and abdomen.Body elongate, straight, slightly flattened dorso-ventrally, more or less parallel-sided, lightly sclerotized, whitish.Integument relatively smooth with scattered simple setae.Body surface without setiferous tubercles, granules, frayed setae or gland openings.Cranium transverse, symmetrical, sclerotized, prognathous, with deep posterior dorsal triangular median emargination, widened posteriorly, partly retracted (except in Micromalthus).Clypeus trapezoidal, usually with unpigmented area posterior to labrum.Median endocarina present, well developed, straight or forked.Epicranial stem and frontal arms not detectable.Hypostomal rods extending into basal half of cranium.Antennae short, not or only slightly protruding beyond level of clypeal apex.Single conical sensorium located ventro-apically on penultimate antennomere.Mandibles short, markedly sclerotized, nearly symmetrical, with relatively wide bases.Mandibles with three apical teeth.Incisor area without teeth or, rarely, with retinaculum (Omma).Stridulatory teeth, prostheca, penicillus, dorsal and ventral carinae, and accessory ventral process absent.Mola markedly developed.Ventral mouthparts slightly retracted.Maxilla with transversely oriented cardo and relatively wide antero-mesally directed stipes.Cardo divided into sclerotized mesal and membranous lateral parts.Two basal maxillary palpomeres subequal in shape, third apical palpomere markedly shorter and narrower.Galea and lacinia subequal or lacinia slightly shorter, fused at base and separated apically, both fused to stipes.Galea with group of setae apically, lacinia with numerous articulated spines and setae apically, and along mesal side.Medial surface of lacinia flattened, and delimited dorsally and ventrally by lines of stout setae.Labium with mentum, prementum and postmentum fused, and constricted between maxillary grooves.Prementum with large wedge-shaped sclerotized ligular sclerome extending apically beyond apices of labial palp.Dorsal surface of ligular sclerome co-joined with flat sclerotized surface of hypopharynx.Labial palps 2-segmented (1-segmented in Micromalthus and first instar of Priacma), widely separated, anteriorly divergent.Labrum transverse, clearly separated dorsally from cranium by clypeolabral suture (except first instar Priacma larvae); with setae along straight anterior edge.Epipharynx bearing pores and setae, not clearly delimited posteriorly.Legs normally present (absent in older instars of Micromalthus), widely separated, 6-segmented consisting of coxa, trochanter, femur, tibia, tarsus and normally two claws (one of which is markedly reduced or apparently absent in first instar of Distocupes and Priacma).Abdominal segments I-III combined longer than thorax.Urogomphi absent.Respiratory system of peripneustic type with annular functional spiracles present on mesothorax and abdominal segments I-VIII; thoracic spiracles larger than those on abdomen.Spiracular closing apparatus present.

KEY TO FAMILIES AND SUBFAMILIES OF ARCHOSTEMATA LARVAE
1 Abdominal apex sclerotized, terminated by one (Figs 19,20,34,35,71,76,77) or two (Figs 9,10,64,65) 57): Body evenly narrowing posteriorly.Cranium with median endocarina forked with both apical arms additionally forked; two additional endocarinae present lateral of median line; frontoclypeal suture absent; 4 stemmata present; cranium nearly parallel-sided.Antenna short, slightly extending beyond clypeal apex; with 4 antennomeres, basal antennomere not reduced in size; apical antennomere not longer than half the length of the rest of antenna.Mandibles with retinaculum; dorsal tooth shortest and ventral longest; transverse ridges on mola absent; dorsal surface of left mandible with nonsclerotised short appendage.The shape of mola deformed in the specimen examined.Maxillary palpifer clearly delimited; sensory spot on lateral surface of apical maxillary and labial palpomeres present, medium sized; apical maxillary and labial palpomeres with single palpal sensorium about half as long as respective apical palpomeres; non-articulated apical cuticular projections present mesally on first and second maxillary palpomeres; narrow and anteriorly oriented sclerotized fixed process present on ventral surface of lacinia; dorsal membranous projection with about 20 cuticular non-articulated teeth present on labio-maxillary articulation membrane; ventral surface of ligular sclerome with 2 setae in transverse line; labial palps 2-segmented.Prothorax ventrally without field of asperities; legs present; claws 2, subequal in length; coxa with 1 tooth anteriorly.Lateral bulge on abdomen absent; abdominal segment IX without asperities; abdominal segment IX membranous and rounded, not pointed; anal opening located apically; abdominal segment X reduced, not visible externally.

Diagnosis
First instar (Figs 1-10): like older instars, except for the following: smaller, legs present and fairly long; chaetotaxy different.Other characters are: clypeolabral suture present and labrum clearly separated from cranium; frontal suture absent; dorsal medial emargination of cranium markedly developed, deep; projection of cranium dorsad of antennifer absent; stemma absent (presence of single stemma was reported for the specimens collected in Hong Kong.I studied two larvae from Hong Kong collected in 1956 and found no trace of stemmata; specimens, however, were markedly degraded); antenna with 3 antennomeres, basal one markedly reduced; mandibles with 3 apical teeth and without additional ventral and basal smaller tooth; anterior edge of labrum straight; labial palps with 1 palpomere; thorax and abdominal segments I-VIII without transverse membranous folds dorsally and ventrally; abdomen terminates with tergal and sternal toothed and curved opposite processes.Chaetotaxy: cranium (Figs 1, 2) with setae 1-9, 13-24 and pores A-D, H-P, R; prothorax (Figs 7,8)  Antenna markedly shortened, not extending beyond clypeal apex; antenna with 4 antennomeres, basal one highly reduced and antenna appearing 3-segmented; apical antennomere markedly elongated, about as long as the rest of antenna.Mandibles without retinaculum; dorsal and ventral tooth about same length, middle one longest; transverse ridges present on mola; dorsal surface of left mandible without non-sclerotised short appendage.Maxillary palpifer poorly delimited; sensory spot on lateral surface of apical maxillary and labial palpomeres absent; each apical maxillary and labial palpomeres with one long palpal sensorium as long as respective apical palpomere, which therefore appear subdivided into one wider and one narrower substructure (represented by palpal sensorium); non-articulated apical cuticular projec-tions mesally on first and second maxillary palpomeres absent; narrow and anteriorly oriented sclerotized fixed process on ventral surface of lacinia absent; dorsal membranous projection on labio-maxillary articulation absent; ventral surface of ligular sclerome with 2 setae in transverse line; labial palps 1-segmented.Prothorax ventrally without field of asperities; legs absent.Lateral bulge on abdomen absent; abdominal segment IX without asperities; abdominal segment IX with sclerotized, toothed and curved tergal and sternal processes; anal opening located apically; abdominal segment X not visible externally.
Archostemata larvae are characterised by the peculiar structure of apices of the labial and maxillary palps.In all larvae the penultimate palpomere bears a palpal sensorium, which is represented either by a single relatively large structure (Figs 6,29,40,41,43,62,63) or, in older instar larvae of Cupedinae, by a compact group of smaller and similar sensoria (Figs 85,86,92,100,102,103).I am not aware of similar structures in other Coleopteran or Neuropteran larvae and, therefore, this sensorium on the labial and maxillary palps might be an autapomorphic character for Archostemata.Additionally, the Archostemata larvae studied are characterised by marked similarities in epipharynx, maxillae and antennae, which, however, currently can hardly be put in a phylogenetic context due to the difficulties of distinguishing discrete and independent characters in these structures.Moreover, first instar larvae of Archostemata (Micromalthus, Priacma and Distocupes) have a markedly similar chaetotaxy on cranium, thorax and abdominal segments I-VIII, which also might eventually provide additional autapomorphies for the group (see also below).

Position of Micromalthus
When described, the genus Micromalthus was assigned to Lymexylidae and since then many authors have discussed the taxonomic position or phylogenetic affinities of this remarkable taxon (see Beutel & Hörnschemeyer, 2002a for more details).Forbes (1926) was apparently the first to propose archostematan relationships of Micromalthus based on a study of wing-folding patterns; this view was corroborated by Böving & Craighead (1931) based on larval morphology.This hypothesis dominates in publications of recent authors (Crowson, 1955(Crowson, , 1981;;Lawrence, 1982Lawrence, , 1991;;Lawrence & Newton, 1982, 1995;Kukalová-Peck & Lawrence, 1993;Beutel & Haas, 2000).However, Baehr (in: Hennig, 1981: 308) considered this genus to be a simplified member of Cantharoidea or Lymexyloidea based primarily on the characters of adults.Barlet (1996) corroborated this view giving reasons why Micromalthus is a lymexylid.Recent revision of Lymexylidae by Wheeler (1986) does not treat Micromalthus as a member.The most recent work by Beutel & Hörnschemeyer (2002a) on the larval morphology and anatomy of Micromalthus clearly supports archostematan affinities of the genus and indicates that the family Cupedidae is a sister-group to Micromalthidae, and the present work supports their conclusions.Following features are potential synapomorphies: cranium is posteriorly widened and laterally rounded; number of stemmata is reduced to one or stemmata absent; antennae are markedly shortened and do not extend beyond clypeal apex; see also Beutel & Hörnschemeyer (2002a: 185-186).

Cupedidae larvae
Older instar Cupedinae larvae are remarkably similar.I was unable to provide reliable diagnostic characters to distinguish the genera based on external morphology because of the limited number of specimens.Larvae of Tenomerga and Rhipsideigma, however, differ from those of Distocupes by having a lateral longitudinal bulge on each side of the abdominal segments I-VIII (Figs 105-107; character noted by Beutel & Hörnschemeyer, 2002a), while some Distocupes larvae have more than four antennomeres (Figs 78, 80, 83;Lawrence, 1991).There are differences in number of sub-elements in maxillary palpal sensorium varying within the subfamily from three to 15 (Figs 85,102) with Rhipsideigma having the highest number.Larvae of this genus have most of setae on body, most notably on dorsal surface of cranium (Fig. 97).Otherwise older instar larvae of the subfamily Cupedinae are generally similar.A basal position of Priacma within Cupedidae is suggested by the following presumptive autapomorphic character of Cupedinae: cranium dorsally is without frontal sutures and therefore frontal sclerites are completely fused with parietal sclerites.

Noteworthy morphological characters of Archostemata larvae
All Archostemata larvae with the exception of those of Micromalthus, have a presumably derived character: a sensory spot on lateral surface of apical labial and maxillary palpomere (43,85,86,102,103), which may be a potential synapomorphy of Ommatidae + Cupedidae.
All Archostemata larvae are characterised by having a maximum of two claws, and in older instar Omma and first instar Micromalthus they are of relatively large and equal size (Figs 5,56,57).In older instar larvae of Cupedidae, however, the posterior claw is variably reduced (94,95,(115)(116)(117), from equal to the anterior claw to almost reduced.These different degrees of claw reduction might be seen on different legs of the same larva.First instar larvae of Cupedidae have, apparently, the posterior claw completely reduced and the legs, therefore, appear to have only one claw (Figs 15,30).

Chaetotaxy of first instar larvae of Archostemata.
This paper presents the first attempt to document the diversity of chaetotaxy in first instar Archostemata larvae.Since the chaetotaxy work with first instar Archostemata larvae is hampered by a scarcity of material, the description is restricted to the most easily observed body parts, namely the cranium and body segments (except metanotum).No attempts were made to provide a detailed description of the chaetotaxy of head appendages and legs.No firm homology is postulated between similarly numbered setae and pores on homologous body parts in Micromalthus, Priacma and Distocupes.I believe, however, that the majority of the similarly designated sensillae on cranium, prothorax, mesothorax and abdominal segments I-VIII are indeed homologous (asterisk (*) near sensillar number indicates the most ambiguous cases of homology).The chaetotaxy of abdominal segments IX and X were found to be markedly different and consequently their sensory structures are simply numbered without any presumption of homology.The structure and chaetotaxy of the metathorax is similar to that of the mesothorax with the most notable exception of the absence of the spiracle and associated seta 23.
The general practice in coleopteran chaetotaxy is to establish a generalised reference system for a family (see references in Material and Methods).The reference system should include the maximum number of recognisable sensory elements.This reference system does not necessarily have to be the most plesiotypic set of sensilla 288 similar to that of a larva of a stem species of the group.The only role of this reference system it to name similarly located and presumably homologous sensillae in larvae of related species.Phylogenetic polarisation of differences in chaetotaxy should be done by using an outgroup as in the analysis of Trechitae (Carabidae) larvae (Grebennikov & Maddison, 2004).Consequently, the absence of a given sensilla does not necessarily imply that this is an apomorphic character, as is sometimes believed.Establishing a reference system for chaetotaxy requires a relatively large number of representatives of a given group to be studied in order to find the optimal set of sen-  rium, ventral; 115-117 -claws; 118, 119 -hind leg, anterior (118) and posterior (119); 120 -fore leg, anterior.
sillae for designation.The optimal criteria imply that this reference system should be relatively similar to the larval chaetotaxy patterns of the majority of species within the group.For Archostemata such an approach is currently hardly possible due to the scarcity of material.In terms of the presence versus absence of sensillae, there are no significant differences between the larvae of the three species studied.The location of sensillae in Micromalthus, however, differs markedly from that in Distocupes and Priacma, and, therefore, it is plausible that the Archostemata larval chaetotaxy reference system will be more similar to that found in the latter two taxa.

CONCLUDING REMARKS
Larval morphology strongly suggests that Archostemata is a natural group and that the bizarre Micromalthus is indeed a member.Chaetotaxy of first instar archostematan larvae proved to be an informative source of characters, however more larvae have to be studied.Special efforts should be directed towards obtaining larvae of Tetraphalerus in South America, Sikhotealinia in Russian Far East and Crowsoniella in Italy.This might not be an easy task, since for the latter two taxa the only material are types series, and in the case of Sikhotealinia it is a single beetle.

FAMILY
: like older instars, except for the following: smaller, body almost parallel-sided; palpal sensorium represented by single structure and not by compact group of sensoria; different chaetotaxy.Other