http://www.collembola.org/publicat/unguis.htm - Last updated on 2017.08.31 by Frans Janssens

Checklist of the Collembola: Note on the Morphology and Origin of the Foot of the Collembola

Frans Janssens, Department of Biology, University of Antwerp, Antwerp, B-2020, Belgium

The arthropodan foot

Literature data on hexapod limb attachement devices are sometimes almost useless, because of confusing terminology (Beutel & Gorb, 2001:178). A glossary of limb end terms is provided in appendix. The terms used here are largely in accordance with the definitions given by Dashman (1953, cited from Beutel & Gorb, 2001:178).

In the majority of arthropods the limb ends in a simple claw-like articulation, which in the Crustacea is known as the dactylopodite (Snodgrass, 1935). The crustacean dactylopodite is provided with two muscles, a levator and a depressor, both arising in the propodite. The simple claw-like limb end is a myriapod feature, but one occurring also in phalangids, Protura and in many pterygote larvae.
In most Hexapoda, a simple claw-like end articulation of the limb occurs in the Protura, in the larvae of many Coleoptera, and in the larvae of Lepidoptera and Tenthredinidae. In the Collembola, the distal unguis bearing articulation is not derived from the crustacean dactylopodite but from the propodite.
In the Hexapoda, the claw bearing articulation differs from the crustacean dactylopodite or propodite in lacking a levator muscle and in having the fibers of the depressor muscle distributed in the tibia and the femur.

In most adult, nymphal, and larval insects, the protarsus, the most distal tarsomere, bears a pair of articulated lateral claws situated upon its base and articulated dorsally to the end of the tarsus . The body of the pretarsus is reduced to a small median claw or a lobe-like structure. The median claw is well preserved in the Lepismatidae and the tendon of the depressor muscle arises from the ventral lip of its base. In Japyx, the base of the pretarsus forms a large plate ventrally upon which is attached the depressor tendon, while its tip is reduced to a minute median claw lying dorsally between the bases of the lateral claws.
The typical protarsus, or terminal foot structure, in insects having true lateral claws arises from the end of the tarsus by a membranous base, upon which are supported the pair of lateral claws and a median lobe, the arolium (pretarsus). The lateral claws are hollow multicellular organs and their cavities are continuous with the lumen of the protarsus. Each claw is articulated dorsally to the unguifer, a median process of the distal end of the last tarsomere. On the ventral surface of the pretarsus is a median basal plate, the unguitractor, which is partly invaginated into the end of the tarsus. To its proximal end is attached the tendon-like apodeme of the depressor muscle of the pretarsus, usually called the retractor of the claws. The unguitractor plate may be divided into two sclerites, or sometimes there is a sclerite distal to it distinguished as the planta. Lateral plates beneath the bases of the claws are termed auxiliae. In the Diptera two large lateral lobes of the foot, known as the pulvilli, arise from the auxiliary plates, one beneath the base of each lateral claw, and there is commonly also present a median process, or empodium, arising from the distal end of the unguitractor plate. The empodium may have the form of a spine, or it may be lobe-like and similar in form to the pulvilli.

In most Arachnida, Pycnogonida, and most insects the protarsus is typically armed with a pair of lateral claws. The pretarsus itself is reduced to a median hook or spur, or reduced completely or obliterated, and the protarsus becomes secondarily a two-clawed structure.
The pretarsus of the pycnogonids is a small dactylopodite with levator and depressor muscles arising, as in Xiphosura and Crustacea, in the tarsus.
The crustacean dactylopodite is usually a simple clawlike segment, though it may be opposed by a process of the propodite, forming a chela. In some of the Isopoda, however, the dactylopodite bears a pair of small claws on its base similar to the lateral claws of insects and some arachnids. The dactylopodite is provided with levator and depressor muscles, which arise in the propodite (tarsus).
The symphylan claw-like pretarsus has a small posterior claw arising from its base. The pretarsal musculature, according to H. E. Ewing (1928), consists of a depressor muscle only, the fibers of which, as in Diplopoda, arise in the tibia.
The chilopod pretarsus is a small dactylopodite-like claw; it is provided with a depressor muscle only, the fibers of which arise in the tibia and the femur and are inserted by a long tendon on the ventral edge of the base of the pretarsus.
The hexapod pretarsus in its simplest form consists of a small claw similar to the terminal claw of a chilopod or diplopod limb, and, as in these two groups, it is provided with a depressor muscle only. This muscle arises usually by several branches distributed in the tibia and the femur, which are inserted on a long slender apodeme or tendon that traverses the tarsus to its attachment on the ventral lip of the base of the pretarsus. The usual pretarsus of adult insects comprises a pair of lateral claws, the ungues, articulated dorsally to the end of the tarsus, and a median structure which is probably a remnant of the primary dactylopodite. A condition intermediate between the one-clawed and two-clawed types of structure is found in some of the Thysanura where there are two articulated lateral claws, and a small median claw, to the base of which is attached the tendon of the depressor muscle. In adult pterygote insects the tendon of the depressor muscle, the retractor of the claws, is usually attached to a small ventral sclerite in the base of the protarsus. The lateral claws are clearly secondary structures developed dorsally from the base of the protarsus.

The collembolan thoracic limb end, conventionally called the pretarsus, apparently bears two simple claws, of which the large, upper, outer one, the unguis, is always present and the small, lower, inner one, the unguiculus, is optionally present.

Basic morphology of the foot of Collembola

Fig.bfc. Basic foot components of Hypogastrura
Aquarel by Frans Janssens © 2004
Modified after Hopkin, S.P. 2004

The foot of Collembola comprises a large outer claw, often called 'upper claw', the unguis (fig.bfc: U), and optionally a much smaller inner claw, often called 'lower claw', the unguiculus (u), empodium or empodial appendage, both arising from a distal section of the limb, conventionnaly called the praetarsus (e.g. Absolon, 1903:99) or pretarsus (e.g. Betsch, 1980:56), also called the tarsus (e.g. Fjellberg, 1998:11), but here called the 'pseudotarsus'. The pseudotarsus bears ectad the unguis (U), a large lamellate spine of the ancestral penultimate podomere, and at the basis of the unguis a pair of setae (s), one anteriorly, one posteriorly.
In Oncopodura, the pseudotarsal setae are unequal in length (Szeptycki, 1977:50). In Bourletiellidae, the posterior pseudotarsal seta is lacking (Betsch, 1980:56), an autapomorphy. In Rastriopes, the anterior seta is reduced to a small immovable spine (Betsch, 1980:56). In Stenognathriopes, both pseudotarsal setae are lacking (Betsch, 1980:56).
Entad, the pseudotarsus bears the unguicular tubercle (ut), the remnant ancestral apical podomere, optionally anteriorly extended with the lamellate unguiculus (u). The pseudotarsus flexes entad relative to the tibia (sometimes called the tarsus, cfr Absolon, 1903:99) by rotation via an ectad ball-and-socket (b+s) monocondyle. In this way, the pseudotarsus rotates, in telescopic fashion, partly into the tibial apex. The pseudotarsus basally bears entad a protruding transversal thickening, the pseudotarsal ridge (ptr), sometimes prolonged with papilae (such as in Oncopodura jugoslavica Absolon & Kseneman, 1932:8,figs.6-8) that corresponds with an apical tibial thickening, the tibial ridge (tr). Both ridges form a mechanism that constrains the entad flexing of the pseudotarsus. The unguis often has a pair of lateral teeth and one or more teeth along the inner edge. The unguiculus often terminates in a long apical filament and may have up to three distinct basal lamellae (lam). In many collembolan groups, the unguiculus is reduced, sometimes even absent, such as in Podura aquatica. In subterranean cave forms, the claws tend to have strongly developed teeth and lamellae, probably an adaptation for walking on wet surfaces.

Fig.fep. Epidermal alignement of the foot components in Dicyrtomina ornata
Janssens, F. © 2006

The unguis is universally present (Christiansen & Bellinger, 1980:15). It has a convex outer side and is tapering. Commonly, the unguis is variously lamellate, with free teeth on the lamellae (lateral and inner). In some poduromorphs and isotomids the unguis is almost featureless. In entomobryids there are two inner tooth-bearing lamellae. In certain isotomids and sminthurids there is an outer inflated sheath or tunica, and/or lateral basal toothed lamellae or pseudonychia.

The unguis is a relatively large multicellular process, a hollow spine-like outgrowth of the body wall and therefore lined by a layer of formative epidermal cells. Most of such processes are solidly fixed to the surrounding cuticula, but some are movable. The immovable processes are specifically termed spines, the movable ones are distinguished as spurs. For example, in this context, the unguis of Collembola is a highly modified spine. The lateral ungues of the feet of insects are large spurs or modified setae. Therefore, the unguis of Collembola and the ungues of Insecta are not homologous and cannot be compared evolutionary.

The following components of the foot are internally ligned with an epidermal layer: the pseudotarsal body, the ungual body and the unguicular tubercle. The colouration of the foot is due to the pigment granules of the epidermal cells. (fig.fep: specimen of Dicyrtomina ornata from Belgium, ..., legit Agnes ..., 2006.01.12). The ungual lamellae and the unguiculus are transparent cuticular extensions.

Anatomical topology of the foot components

With respect to the topology of the foot of Collembola, we have adopted here the terminology as used by Absolon & Kseneman, who have published an exhaustive discussion on the usage of the topological jargon as used by many authors (1932:5-8). In summary: the convex side of the unguis is considered to be at the outer side of the limb, while the concave edge of the unguis is the inner side. Therefore, we prefer to use the vernacular terms 'outer claw' and 'inner claw' for unguis and unguiculus, respectively.

With respect to the lateral sides of the limbs there is no consensus among authors in the terminology used; even by thesame author, inconsistent terminology is used; due to the different position of the limb in relation to the body, and due to the asymmetric topology of the foot components, thesame lateral side sometimes is refered to as the outer side and sometimes the inner side (Absolon & Kseneman, 1932:7-8). Taking into account that the feet of the limbs in relation to the anterio-posterior body axis are topologically different:
in the foot of the prolimb, the unguiculus and the inner ungual edge are slightly caudad;
in the foot of the mesolimb, the unguiculus and the inner ungual edge are entad;
in the foot of the metalimb, the unguiculus and the inner ungual edge are slightly cephalad;
the specimen's body has to be thought of as being in the standard anatomical body position, and
the body side (left/right), the type of limb (pro-, meso-, meta-), and its anterior/posterior aspect should be indicated to unambiguously describe the foot components. Such a system was already in use by Willem (1900:127).

Types of ungual transversal sections

Fig.unglam. Ungual lamellae
Dicyrtomina ornata from the UK
Right mesoleg, frontal aspect
2007 © Valentine, B.

Fig.section. Transversal section of unguis.
2005 © Janssens, F.

Fig.types. Types of ungual transversal sections.
Frans Janssens © 2005.

Comparing transversal sections of ungues, several types of sections can be distinguished: bicuspidate, tricuspidate, quadricuspidate and cinqocuspidate (fig.types). Each section is characterised by a convex outer ungual body wall. Both the concave inner ungual body walls of the tricuspidate section are joined medially forming as such the inner edge of the unguis. The bicuspidate section has only one concave inner body wall, interconnecting the anterior and posterior lateral edges; a medial inner edge is therefore lacking.
All sections can be derived from the tricuspidate one:
in the quadricuspidate section, the median inner edge is split, resulting in an anterior and posterior inner edge;
in the cinqocuspidate section, the inner edge is partly sunk into the ungual body, resulting in addition to the medial inner edge in an anterior and posterior rim (rounded edge);
in the bicuspidate section, the median inner edge is reduced completely.

Poduromorpha, typically have an unguis with a tricuspidate transversal section. This type of unguis is the plesiomorph type. It can also be found in the entomobryomorph Actaletidae, Coenaletidae, Tomoceridae, Oncopoduridae, Isotomidae, Microfalculidae and in the Symphypleona. The section of the unguis of the entomobryomorph Entomobryidae, Cyphoderidae and Paronellidae is proximally quadricuspidate, while more apically it is tricuspidate. In some cave entomobryids, the apical inner edge is completely reduced. In those species, the apical transversal section of the unguis is bicuspidate. That entomobryomorph Collembola have two types of ungues might suggest that Entomobryomorpha is an artificial groupment. The transversal section of the unguis of Neelipleona is in principal tricuspidate. Neelus and Megalothorax have proximally a cinqocuspidate section.

Model of the unguis of Poduromorpha

Fig.4. Basic form of the poduromorph unguis.
Frans Janssens © 1999.

" Die an den Tarsen insierten Klauen sind von sehr einfacher und ursprünglicher Gestalt. Die sogenannte obere Klaue ist von mehr oder weniger langer, meist etwas gekrümmter Form, im Querschnitt 3kantig, indem 2 Lateralkanten (rechts und links) und eine Innenkante ausgebildet ist. Die Lateralkanten sind nach ausen unter einander durch eine mehr oder minder konvexe Fläche, mit der Innenkante je durch eine meist stark konkave Fläche verbunden. Die an den Lateralkanten stets beiderseits auftretenden Zähne bezeichne ich als Lateralzähne, die auf der Innenkante insierten als Innenzähne. Weder Aussenzähne noch eine Tunica finden sich in dieser Familie [Achorutidae]. " (Börner, 1901:13-14).

In Poduromorpha, the unguis, in profile, has the shape of a blade of a knife slightly curved entad, with a strong basis and pointed apex; in transversal section, one distinguishes one inner edge and two lateral edges (Thibaud, 1970:119). The habitus of the unguis in its most basic form, is a modified polyhedron (Fig.4): a pyramid with a trianguloid base and with three three-dimensionally curved trigonal faces: one outer face and two inner lateral faces. Its geometrical shape is an irregular tetrahedroid with four vertices, six edges and four not identical triangular faces.
The unguis is enantiomorph, bilaterally symmetrical. Laterally viewed (Fig.4, left and right), the unguis is proximodistally curved entad: the outer face is convex and the two inner lateral faces are concave; it is tapering gradually from its basis at the pseudotarsus to its apex, where the three faces join in a sharp point.

Foot of Podura aquatica.
Frans Janssens © 1999.

The base of the unguis or a transversal cross-section of the unguis is a tricuspoidal face with an outer convex edge and two inner concave edges. The proximodistal intersection of the two lateral patches forms a more or less sharp median edge. From an inner view, the lateral edges are sligthly convex (Fig.4, middle).

Basic geometric characteristics of the unguis:

• length,
• width of basis,
• radius of the proximodistal curvature of the outer face,
• radius of the transversal curvature of the outer face,
• radius of the proximodistal curvature of the inner lateral faces,
• radius of the transversal curvature of the inner lateral faces.

Fig.5. Poduromorph unguis with fingernail-like outer sclerite.
Frans Janssens © 1999.

The ungual outer face is often enforced with a fingernail-like sclerite without or with lateral teeth (basally and/or apically) (Fig.5). This outer sclerite can be weak to strong, to very strong as in Axenyllodes echinatus FJELLBERG, 1988 [Fjellberg, 1998:Fig.76C] and can cover the outer face up to the basis of the unguis (Fig.5, left) or just partly, leaving a basal portion of the unguis uncovered (Fig.5, right) as in many Onychiuridae.
Also the ungual inner edge is often enforced.

The pseudotarsus (Fig.3) and the lateral faces of the unguis are typically provided with the epicuticular ultrastructure of microtubercles, as in Folsomia candida (Thibaud, 1970:119). The microtubercles are interconnected by epicuticular hollow ridges [Hopkin, 1997:55] in a geometric arrangement. The superfacial geometric texture increases the strength and - more specifically - the rigidity of the cuticle.

Fig.6. Poduromorph unguis with inner tooth.
Frans Janssens © 1999.

The ungual inner edge is often armed with a tooth; this tooth is situated subapically such as in Hypogastrura, Schoetella, Mesogastrura, Mesachorutes, Acherontiella, Xenylla, and medially such as in Ceratophysella, Schaefferia, Typhlogastrura (Thibaud, 1970:119). The lateral edges may have small teeth, basally situated, such as in many Hypogastruridae (Thibaud, 1970:119) or medially, such as in Anurida maritima.

The unguis of Typhlogastrura, Schaefferia and some Ceratophysella are more slender, elongate (much longer then wide), and longer then the tibia, then in other Hypogastruridae; e.g. in Ceratophysella bengtssoni, the length of the unguis = 35-40 micron, tibia = 50-55 micron. (Thibaud, 1970:119). List of podumorph ungual features (derived from [Fjellberg, 1998]):

• length: long; as long as tibia
• shape: broad, elongated, narrow, slender; unarmed, without visible teeth; simple; with neck-like basal portion, with a neck-like constriction; with strong outer 'nail'; expanded shovel-like;
• inner tooth: absent (Fig.6, middle), or if present:
  • at 1/3 from apex (Fig.6, right); near mid, at mid, at middle, near middle (Fig.6, left),
  • single,
  • weak, small (Fig.6, right); strong, large, distinct (Fig.6, left); variable,
• lateral teeth: absent, or if present:
  • near base, near apex, at 1/3 from base, at base, in basal part,
  • small, distinct, invisible but may be present, indistinct, weaker, strong, vestigial, variable, pointed, basal,
  • one pair, two pairs,
• apical part: thin, transparant.

Fig.7. Model of unguis of Anurida maritima.
Frans Janssens © 2002.

We conclude that in the dominant foot position, the ungual tip, which is much more flexible compared to the ungual body, bends entad in a convex way to adapt to the the substrate surface form while the lateral teeth penetrate the substrate and/or its texture. Therefore, the function of the lateral ungual teeth is to improve traction during locomotion. The stiffness of the ungual tip is dependant on the presence of the median lamella. An absent median lamella makes the ungual tip most flexible. The larger the median lamella, the stiffer the ungual tip. We conclude that the function of the ungual tooth of the median lamella is to improve the bendability of the ungual tip even if a large median lamella is present. While the ungual tip bends entad, the median tooth closes. When closed entirely, the tip will not bend any further.

Fig.H20-1. Force system on poduromorph unguis at water surface.
Frans Janssens © 2004.

Walking on water. The outer face of the unguis is smooth and hydrophile; the inner lateral faces are provided with the hydrophobe and plastron bearing epicuticular ornamentation; when penetrating the water surface, the unguis is wetted by the water surface film at its outer face, creating a concave meniscus, firmly adhering the unguis to the water surface, while the inner lateral faces repel the surface water film, creating each a convex meniscus at both inner sides of the unguis; these constitutes a steep dip in the water surface in which the inner edge of the unguis functions as an orthogonal position stabilisator; the inner edge tooth prevents too deep penetration of the unguis in the water; in addition, the unguiculus and tenent hairs support stabilising the foot on the water surface (modified after Thibaud, 1970:182).

Fig.H20-2. Orthogonal stabilisation of the poduromorph unguis at the water surface.
Frans Janssens © 2004.

The unguis, when penetrating the water surface, is subjected to several forces that reach an equilibrium at a given point (fig.H20-1). Due to the surface tension forces, applied in opposite directions to the outer face versus the inner lateral faces, the foot is firmly anchored at the water surface film level. The weight applied to the foot (G), in combination with the gravitad components of the water surface tension forces applied to the outer ungual face (Fa1 and Fp1), are balanced against the water surface tension repellent force (F): F = G + Fa1 + Fp1. The laterally applied surface tension forces (Fa2 and Fp2) warrant a stable positioning of the foot on the water surface (Fa2 = Fp2).
When the unguis is not in a perpendicular position to the water surface, the menisci at the lateral edges are unequal (fig.H20-2). The larger meniscus will apply a larger lateral force (Fa2) to the unguis than the smaller meniscus (Fp2). Due to the larger volume of water, the gravity point of the larger meniscus is more distant from the water surface level than that of the smaller meniscus at the opposite side of the unguis. As such, Fa2 and Fp2 apply a momentum to the unguis that will try to correct the oblique position of the unguis. In this way the surface tension acts as an ungual self-positioning system.

The foot of Hypogastruridae is well adapted to walk on water or on a wet substrate: the unguis bears a teeth, which is situated on the inner edge, medially (such as in Ceratophysella, Schaefferia, Typhlogastrura) or more apically (such as in Hypogastrura, Schoettella, Ceratophysella, Mesogastrura, Mesachorutes, Acherontiella, Xenylla); especially in the former position, the teeth constitutes a more efficient system to assure the flottation; in Typhlogastrura and Schaefferia, the unguis is relativelly slender and longer, which facilitates in this way the walking on water (Thibaud, 1970:182).

Model of the unguis of Entomobryomorpha

" Zum Verständnis der Verwandtschaft der einzelnen Gruppen, insbesondere der Isotomini, Tomocerini und Entomobryini, ist der Bau der oberen Klaue besonders wichtig. Wie ich bereits in meiner zweiten vorläufigen Mitteilung (8) auseinandergesetzt habe, sind die Isotomini und Tomocerini die Vertreter des einfachsten Typus; bei ihnen ist die obere Klaue ein in Querschnitt dreikantiges Gebilde mit einfacher, d.h. ungespaltener Innenkante, auf der vor- resp. hintereinander, niemals neben einander inserierte Zähne stehen können. V. Willem (33) giebt für Tomocerus in seiner kürzlich erschienenem Monographie an, dass die Innenkante mit Doppelzähnen bewaffnet sei, etwa so, wie ich es für die Proximalzähne der Innenkante der Entomobryini beschrieben habe. Er stellt seine Beschreibung der alten von Tullberg gegenüber und erklärt die letztere für falsch. Nachdem ich daraufhin abermals diese Verhältnisse untersucht habe, kann ich nur meine eigenen früheren Angaben bestätigen, die volkommen in Übereinstimmung mit denen von Tullberg stehen. So muss ich denn auch die betreffende Figur Willem's (Tafel IX,6,7) als unrichtig bezeichnen die thatsächlich vorliegenden Verhältnisse findet man in Fig.15 für Tomocerus plumbeus (L.) Tullb. abgebildet.
Andere Verhältnisse zeigt uns die Innenkante der oberen Klaue bei sämtlichen Entomobryini (Fig.16 etc.). Bei diesen ist die Innenkante über der Basis gespalten, sodass der Proximalzahn der Tomocerini[(sic)] doppelt wird, was man deutlich in der Figur 16 erkennen kann. Der nächste distale Zahn ist aber niemals mehr doppelt, sodern im Gegensatz zu den Abbildungen Willems, einfach, wie bei den Tomocerini und Isotomini. " (Börner, 1901:39-40).
" Unterfamilie: Tomocerini(sic) Schäffer, Börner.
Obere Klaue mit einfacher, ungespaltener Innenkante. " (Börner, 1901:60).
" Unterfamilie: Entomobryini(sic) Schäffer, Börner.
Innenkante der oberen Klaue an der Basis gespalten. " (Börner, 1901:61).
" Entomobryaeformes: die Zweispaltigkeit der Innenkante der oberen Klaue ist nur sehr schwer zu erkennen, da die beiden Teilhäften sehr nahe bei einander und die auf ihnen stehenden Zähne stets genau neben einander liegen.
Lepidocyrtiformes: die Zweispaltigkeit der Innenkante der oberen Klaue ist sehr leicht zu erkennen, da einmal die auf den Teilkanten stehenden Zähne meist gross und ferner häufig nicht unmittelbar neben-, sondern oft etwas vor-, resp. hinter einander stehen. " (Börner, 1901:62).
" Gattung Orchesella Templ.
Betreffs der oberen Klaue möchte ich noch bemerken, dass sich ausser den Lateralzähnen auch 2 echte Aussenzähne nahe der Basis vorfinden, die den übrigen Entomobryiden fehlen; 1 solcher findet sich in der Gattung Lepidocyrtus Bourl. " (Börner, 1901:65).
" Cyphoderus albinos Nic.
Wie ich schon oben in der Gattungsübersicht anführte, besitzt die obere Klaue einen ganz anderen Bau als man bisher angenommen hat. Die von Tullberg (31) gegebene Figur (Tafel VI, Fig.17) giebt nicht nur die obere, sondern auch die untere Klaue unrichtig wieder. Die Innenkante der obere Klaue ist fast bis zur Mitte gespalten, etwas vor der Mitte befindet sich auf der internen Teilkante ein grosser Zahn, der fast die Länge der unteren Klaue erreicht (Fig.28). " (Börner, 1901:71).

" Willem gibt überhaupt, wie für Tomocerus, so auch für Orchesella an, daß an der Innenlamelle der Klaue Doppelzähne vorkommen, wie er aus auf der Taf.IX, Fig.6,7 und Taf.X,Fig.3 seines sub Note 2 zitierten Werkes abbildet. Ich habe schon frürher selbständig dieses Thema studiert und muß heute nur die bezügliche Börnersche Korrektur bestätigen. Tomocerus besitzt überhaupt nur einfache Zähne und bei Orchesella (als Prototypen der Subfamilien betrachtet) ist nur der Proximalzahn doppelt. Gerade dieser Unterschied im Baue des ersten proximalen Zahnes ist sehr wichtig, indem auch Tritomurus und, wie ich weiter zeigen werde, auch Lepidophorella durch dieses Merkmal im Gegensatze zu allen Entomobryini stehen.
In einem anderen Punkte kan ich aber Börner nicht zustimmen. Er spricht bei Tomocerus von einer einfachen, das ist "ungespaltenen Innenkante", bei Orchesella von einer "über der Basis gespaltenen Innenkante". Meine Untersuchungen führen zu dem Resultate, daß die ventrale Lamelle bei beiden genannten Gattungen (als Prototypen) gleich gebaut ist. Sie soll eigentlich als Doppellamelle bezeichnet werden, denn beide Kanten sind an der Naht, an der eigentlichen ventralen Lamelle verwachsen. Nur bei Orchesella (als Prototyp) besitzt jede einzelne Lamelle ihren eigenen Proximalzahn, wogegen alle übrigen Zähne beiden Kanten gemeinschaftlich sind. Wir können diese Verhältnisse namentlich auf der Tomocerus- und Orchesella-Klaue gut beobachten, wenn wir gleichzeitig diese Klaue lateral und ventral untersuchen.
Die unbedeutende Größe, ungünstige Lage und teilweise auch die Durchsichtigkeit der appendiculären Teile, namentlich der sogenannten Doppelklaue und der Mucrones, erschwert sehr eine genaue mikroskopische Untersuchung, so daß in den betreffenden Angaben der Autoren bedeutende Undeutlichkeit herrscht, die sich namentlich in ihren Figuren gut kennzeichnet, wo einzelne Lamellen und Kanten unrichtig gezeichnet, verbunden und verwechselt werden. Gewöhnlich sehen wir die laterale Kante mit der ventralen Lamelle ein Dreieck bilden (siehe z.B. Schött: Zur Systematik und Verbreitung paläarktischer Collembola [1893], Taf.III,Fig.13; Taf.IV,Fig.7; Taf.VI,Fig.6,8,33 etc.). Es läßt sich dann auf eine enorme Verschiedenheit in der Form der Klaue schließen. Und doch ist die Vermutung des Dr. J.C.H. de Meijere [Über das letzte Glied der Beine bei den Arthropoden, Zool. Jahrb., Bd XIV, Heft 3, 1901] ganz richtig, die er mit diesen Worten äußert: "Obzwar ich nicht viele Collembolen untersucht habe, scheint es mir doch sehr unwahrscheinlich, daß darunter (das ist im Baue der Klaue) so sehr verschiedene Verhältnisse vorkommen werden, wie die Abbildungen vermuten lassen." Die Sache verhält sich tatsächlich so; die Klaue ist namentlich bei den Arthropleona nach demselben Prinzip gebaut und wenn wir uns einer einheitlichen Terminologie für alle Kanten, Lamellen, Zähne etc. anschließen, wird dadurch ein sehr schwieriger Abschnitt bei einer wissenschaftlichen Bestimmung der Collembola-Arten und -Gattungen erleichtert. Es genügt, die schon usuellen, von Tullberg, Willem und Börner eingeführten Termini zu ergaänzen, in einem Falle vielleicht zu ändern.
Wenn wir die Klaue lateral beobachten (Taf.I,Fig.5; Taf.II,Fig.1,8,13), so erblicken wir gewöhnlich fünf nebeneinander von der Spitze verlaufende "Linien". Die erste ist die dorsale Linie der Klaue (d.), die zweite die obere laterale Kante (lk.1), die dritte (gwöhnlich undeutlich) die durchschimmernde untere laterale Kante (lk.2), die vierte die ventrale Lamelle (v.l.) (nach Börner Innenkante) und endlich die fünfte die hintereinander liegende Kanten der ventralen Lamelle (v.k.1, v.k.2). An der fünften "Linie" sitzen die Zähne und basal können wir ganz gut beobachten, daß sich da eigentlich zwei "Linien" (siehe Taf.II,Fig.1) ziehen, die vierte Linie (v.l.) kreuzen und dann die Naht zwischen Praetarsus (pt.) und Klaue bilden, wie es auch Börner in seiner Fig.16 (oben S.101, N.6) sehr gut, in Fig.15 undeutlich zeichnet. Wenn wir dann die Klaue ventral beobachten (Taf.I,Fig.6, Taf.II,Fig.2), so verstehen wir das ganze gleich und leicht. An beiden Seiten ziehen sich die lateralen Kanten (lk.1, lk.2), von welchen früher lk.1 oben, lk.2 unten lag (vgl. gleichzeitig Taf.I, Fig.5 und 6; Taf.II, Fig.1 und 2), die dorsale Linie (d.) verschwindet natürlich, der Verlauf der ventralen Lamelle (vl.) und ihrer Kanten (vk.1, vk.2) erscheint nach Entfernung des Empodialanhanges so, wie es in Taf.II, Fig.2; Taf.I, Fig.6 abgebildet ist. Laterale Kanten tragen gewöhnlich große Zähne, Pseudonychien oder besser nach Börner laterale Zähne. Die sind gewöhnlich einfach gebaut, glatt, mit zwei Kanten, von welchen die eine kürzere als interne (i.ps.k.), die zweite längere, die laterale Kante der Klaue vertretende Kante, als externe Pseudonychienkante (e.ps.k.) zu bezeichnen ist. Pseudonychien bei Tomocerus (bei allen?) und bei Lepidophorella (konträr: nur bei diesen?) besitzen noch eine mediane, starke, kammartige Lamelle, auf welche Willem zuerst aufmerksam machte und welche er in Fig.7, Pl.IX ganz richtig abbildete. Bei lateralem Anblicke sehen wir also (Taf.II, Fig.1,8) drei Linien, die erste ist die dorsale Linie der Pseudonychie, die zweite ist die externe laterale Kante (e.ps.k.), die dritte die mediane Pseudonychienlamelle (m.ps.l.). Die interne laterale Kante ist nicht sichtbar. Bei den Formen mit einfachen Pseudonychien sehen wir lateral nur zwei Linien (Taf.I, Fig.5; Taf.II, Fig.13), die dorsale Linie und die externe laterale Kante (e.ps.k.).
Nur (soweit bekannt) für die Tomocerinenklaue sind Falten charakteristisch, die sich ventral 3+3 oder 4+4 an der Klauenfläche verbreiten. In der Fig.7 Willems sind sie nicht richtig angegeben. Das sind diejenigen Gebilde, welche früher Anlaß gaben zur Zeichnung von eigentümlichsten, welligen Linien. (Vgl. z.B. K.Absolon: Über einige teils neue Collembolen aus den Höhlen Frankreichs und des südlichen Karstes, Fig.10; H.Schött: Zur Systematik und Verbreitung der paläarktischen Collembola, Taf. III, Fig.8; Folsom: Papers from the Harriman Alaska Expeditions, Apterygota, Pl.VIII, Fig.46,49 etc.). " (Absolon, 1903:102-103).

In Entomobryomorpha, one can distinguish two types of unguis: Tomoceridae, Oncopoduridae, Isotomidae and Microfalculidae have a tricuspidate unguis, while Entomobryidae, Cyphoderidae and Paronellidae have a quadricuspidate unguis. This might suggest that Entomobryomorpha do not form a monophyletic grouping.

Fig.3. Foot of Folsomia.
Hans Henderickx © 1999.

See Fig.3 (left metafoot of Folsomia candida WILLEM, 1902; lateral outside view; base of the pseudotarsus is about 20 micron) for an example of a weak nail-like texture at the outer side of the unguis in Isotomidae. More SEMs of the claws of the cave isotomid Folsomia candida WILLEM, 1902 made by Hans Henderickx © 1999:

• foot complex habitus; lateral inside view
• left proclaw; outer view; note the narrow and elongate shape, typical for stage III cave forms
• left mesoclaw; outer view; close-up of the lateral setae of the pseudotarsus and the concave bilamellate unguiculus
• right proclaw; outer view; close-up of the peculiar shape of the bilamellate unguiculus
• close-up of the cuticula of the ungual lateral faces with the hexagonal pattern of microtubercles
• lateral tooth artefact caused by a dust particle? Note the very small inner proximodistal lamella.

In the monogenic Microfalculidae, the unguis of Microfalcula is stronlgy reduced in size and the lateral edges are greatly broadened (Betsch & Massoud, 1968:907+fig.6G). It appears that the function of the unguis is taken over by the much larger and highly modified 'tenent hair'. According to Betsch & Massoud, (1968:907), 'the tenent hair is migrated to the pretars' ('l'ergot, qui a migré sur le prétarse, ...'). However, this is contradicted by their fig.6A in which the 'tenent hair' clearly originates on the tibia and it is confusing in their fig.6D-G, in which it appears to be part either of the tibia (fig.6D,E,G) or of the pseudotarsus (fig.6F). Betsch & Massoud, (1973:6,fig.C-E) consistently draw the 'tenent hair' as making part of the pseudotarsus. Therefore, we propose an alternative interpretation which is more in line with the groundplan of the foot of Collembola: given the 'tenent hair' makes part of the pseudotarsus, noting that pseudotarsal setae are lacking in the description and figures, and noting that the lateral position of this seta corresponds with the lateral pseudotarsal setae in the groundplan, we conclude that the so-called tenent hair is not a tenent hair but it is a highly modified pseudotarsal seta.

The unguis of entomobryids is quadrilamellate and the inner pair of lamellae have several teeth [Christiansen, 1966:530]. Due to the flexibility of the pseudotarsal integument, the unguis can bend in a forward position (observed in unmodified non-cave stage I forms), a sideways tilted position (observed in intermediately modified stage II forms), and a backward position (observed in highly modified cave stage III forms) depending on the angle of contact and texture of the substrate.
In cave forms, the stage I position is the least effective on smooth wet surfaces and it tends to disappear entirely.
The stage II position is more effective provided the outer lamella is thin enough and the inner lamella large enough to permit contact of its longest region (the basal tooth). An enlargement of the basal tooth serves to increase traction during locomotion. Also the outer lamella tends to elongate and enlarge. Both the elongation and broadening of the outer lamella improve both the penetration and traction on wet surfaces.
In the stage III position, the inner lamella tends to reduce, especially apically, resulting in a long, flat, flexible ungual tip. The basal tooth ceases to function and it tends to regress. When the unguis assumes the standard position it improves locomotion on wet surfaces by penetration. Note that Christiansen (1966:534,536) pointed out that in this stage the inner lamella of the unguiculus redevelops into a rounded form to control the water surface penetration. See Fig.3 for an example of such a convergent evolution in cave Isotomidae.

Tab.3. Characteristics of entomobryid unguis.

Ungual feature	unmodified	intermediate	highly modified
Size	normal	larger/longer	large/elongate
Basis of dorsal lamella	thin	broad	broader
Proximodistal radius of dorsal lamella	large	smaller	smallest
Transversal radius of dorsal lamella	round	flatter	flat
Size inner lamellae	large	large (asymmetric?)	small
Size basal teeth	normal	large	small; vestigial
Position of basal teeth	normal	less basally	more basally

Fig.8. Schematic model of the entomobryid unguis.
Frans Janssens © 1999.

Schematic three-dimensional model of the idealised morphology of the entomobryid unguis (Fig.8):

• outer lateral lamellae (green) with lateral teeth;
• inner basal lamellae (red) with basal teeth;
• longitudinal median inner lamella (yellow), interconnecting the lateral and basal lamellae, with inner tooth and apical tooth.

Fig. bu. Broken unguis of Orchesella flavescens showing nested structure of inner and outer lamellae Frans Janssens © 2004.

Fig. su. Split unguis of right metafoot of Entomobrya nivalis specimen from Belgium, Beveren 2003.10.23, leg. Bruers, J. Frans Janssens © 2004.

In specimens where the unguis accidentally is broken transversally, it can be observed that the inner basal lamellae (see arrow) are actually embedded into the unguis (fig. bu: Orchesella flavescens, left mesofoot, outer aspect, phasecontrast oc.10x obj.100x, immersion oil, negative image; note also the lateral pseudonychia). The inner lamellae appear to be secondary reinforcements nested inside the primary outer lamellae. Is the entomobryid unguis a complex of two elongated parallel encapsulated primary (poduromorph-like) ungues?

Exceptionally, it can be observed that the unguis is split latero-proximodistally (fig. su: Entomobrya nivalis, right metafoot, posterior aspect, phasecontrast oc.10x obj.100x, immersion oil, negative image). It appears that the inner teeth bearing lamellae are formed structurally independent from the outer lamellae. This observation complies with our model of the entomobryid unguis being a compound of two nested substructures. In the split unguis it is also observed that the prolonged apex of the inner substructure, the inner edge, is a structural part of it. When the interconnecting tissue between outer and inner lamellae is lacking for some reason, the lamellae open apically due to the elastic strain of the ungual basal intugement. In other words, in a normal unguis, the inner lamellae apply a permanent tension to the outer lamellae, 'trying to pull themself away'. This tension will keep the tip of the outer lamellae in a continuously bent condition, increasing in this way the proximodistal rigidity of the unguis.

The unguis of Mesentotoma bears from two to four ventral teeth (Christiansen, 1956:15). The basal pair enlarged, often basally joined, and usually heavily reinforced.

In Pogonognathellus, distally, the tibia has entad a bulged process, the tibial ridge, that pushes against the pseudotarsal ridge when the foot is placed on the substrate. In this way the unguiculus is apressed against the unguis. The same happens when the unguis is flexed entad.

Model of the unguis of Symphypleona

" Die obere Klaue stellt in den meisten und ursprünglichen Fällen - wie bei den übrigen Collembolen - ein mehr oder weniger gekrümmtes, nach vorn zu spitz werdendes 3 kantiges Gebilde vor, an dessen seitlichen oder Lateralkanten sogenannte Lateralzähne, an dessen Innenkante Innenzähne auftreten können.
Es kann nun aber eintreten, so zeigt es uns z.B. Sminthurus variegatus Tullb. und Papirius fuscus Lubb., dass die über der Basis stehenden Lateralzähne eine bedeutende Vergrösserung und Differenzierung erleiden, sodass sie uns wie feingezähnte dünne Blätter, sogenannte Pseudonychien erscheinen, welche sich seitlich an die obere Klaue anlegen (Fig.37). Zugleich bemerken wir auf der Aussenseite der oberen Klaue, die von den Lateralkanten eingeschlossen ist, dachziegelartig über einander liegende, niedrige Zähne wie man sie auch bei anderen Sminthurus-Arten (S. viridis (L.) Lubb., S. aquaticus (Bourl.) und Papirius fuscus beobachten kann. Ich möchte diese Aussenzähne als die Vorläufer der Tunica der oberen Klaue auffassen; zu dieser Annahme führte mich u.a. das Vorkommen dieser Aussenzähne in Gemeinschaft mit einer Tunica bei Sminthurus viridis Lubb., S. marginatus Schött und S. flaviceps Tullb., wie auch Papirius minutus (O. Fabr.) Tullb. Über die ontogenetische Entstehung der Tunica ist bisher nichts bekannt geworden.
Die Tunica selbst tritt in sehr verschiedener Form bei den einzelnen Vertretern auf. Sie liegt bald der Klaue eng an, bald steht sie mehr oder minder weit ab. In ersterem Falle ist sie nur schwer zu erkennen, wir aber durch Einwerkung von Kalilauge einmal von der Klaue abgehoben und meist auch etwas gewellt, was ihre häutige Natur sofort hervortreten lässt. So finden wir es bei S. viridis und S. marginatus. Die Tunica erstreckt sich hier bis fast oder ganz an die Spitze der Klaue und verschmilzt mit dieser an den Seitenkanten. Eine absthehende Tunica tritt uns hingegen bei S. fuscus (L.), S. lubbocki Tullb., S. flaviceps Tullb., wie ach bei Papirius minutus (O. Fabr.) Tullb., P. flavosignatus Tullb. und P. dorsalis Reuter entgegen, die der Klaue ein plumpes Aussehen verleiht. Sie hebt sich bei diesen Arten schon dicht über der Basis der Klaue deutlich von dieser selbst ab, erstreckt sich mehr oder minder weit bis zur Klauenspitze, diese jedoch unbedeckt lassend, um dann ebenfalls an den Lateralkanten in das Chitinskelet der oberen Klaue überzugehen (cf. Tafel II, Fig.10a und b). In Figur 38 habe ich den Tarsus des ersten Fusspaares von S. fuscus (L.) abgebildet, wo man über der Tunica auch deutlich das eine Pseudonychium erkennt, das durch eine äusserst feine, am Aussenrande etwas verdickte Membran mit dem Pseudonychium der andere Seite der Klaue in Verbindung steht. " (Börner, 1901:85-87).

The unguis has a tricuspidate cross section with a sharp inner edge, which sometimes bears teeth, and a broad outer side which is smooth or has teeth, a balloonlike duplication (tunica), or lateral, serrate duplications (pseudonychia) (Bretfeld, 1999:7). The unguis is a slightly bent pyramid with a triangular basis, with one inner edge, and two lateral edges. The outer face of the unguis is convex, and epicuticular microtubercles are absent; the postero-internal and antero-internal face are integumentally granulated except along the inner edge of the unguis. The outer face is hydrophile, while the epicuticular microtubercles on the inner faces are hydrophobe. The lateral edges are often provided with teeth; basal teeth may fuse and hypertrophically form two more or less long lateral lamellae, the pseudonychia, well developed in genera such as Dicyrtomina. Often, the external face is covered by a thin, fragile sheet, the tunica.
In Gisinurus malatestai Dallai, 1970:469, a lateral cavity traverses through the ungual body. (Betsch, 1980:56). The unguis, medially at its anterior side, bears an orifice that connects to the cavity in the body of the claw (Nayrolles, 1993:51).

Fig.tun. Foot of Dicyrtomina sp. with distinct tunica
Krebs, C. © 2010.01.24

Note on the function of the tunica of the Symphypleonal unguis as a self-wetting mechanism: In Symphypleona, several species have an unguis with an outer inflated sheat, a kind of membraneous pocket at the outer side of the unguis: the tunica. On 2003.12.04 we have collected some specimens of Dicyrtomina minuta from rhododendrons in the Hof ter Saksen, Beveren, Belgium. Observing the live specimens with a stereo-microscope with cold light source revealed the following remarkable walking behaviour. The specimens were put in a small container that was closed with a transparant cover. Soon, specimens crawled up to the walls of the container and walked upside down on the smooth surface of the transparant cover of the container. The ventral tube was not used. The ungues were not used to hook themselves to the smooth surface of the cover. It was quite easy to see that the ungues were bent entad, as shown in the dominant foot postion fig. fpA, and that each ungual outer face was in contact with the cover and that each unguis was surrounded by a tiny wet spot. Due to the adhesion of the watery surface film surrounding the feet, specimens could walk easily upside down on the smooth cover. Replacing the cover with a new dry one, did not prevent specimens to walk on it upside down. But soon they stopped walking, then produced a drop of liquid from their mouth, catched the drop with a foot and placed the wetted foot on the surface of the cover. After wetting their feet, they were able to continue walking for a while, after which the foot wetting procedure was repeated. Presumably, it is due to the tunica of the unguis, characteristically to Dicyrtomina species, that specimens can walk on such dry surfaces. The membraneous tunica functions as a compressable container of liquid that is emptied on the surface whenever the foot is pressed against the substrate, producing in this way a local surface film. The surface tension created by the local surface film on the foot outer face produces a firm contact with the substrate and allows them to walk on dry substrates, even upside down. To be functional, the outer side of the unguis has to be hydrophile. When the pressure on the foot releases, the compressed tunica relaxes and reopens. It is then refilled with the water of the surrounding surface film due to capillar effects. In this way the tunica of the foot produces its local water surface film that enhances contact with dry substrates. Note that Dicyrtomina species in general live in relative dry habitats, such as vegetation. Other Dicyrtomidae, that lack a tunica, live in more moist habitats where water surface films are readily available and a tunica is not advantadgeous. Therefore, the tunica must be an adaptation for improving contact with the surface when walking on unwetted substrates.

Model of the unguis of Neelipleona

In principal, the transversal section of the unguis of Neelipleona is tricuspidate. In Neelides folsomi, the ungues of the prolimbs and mesolimbs are longer than those of the metalimbs; they are provided with a well developed tooth at the basis of each lateral lamella (Börner, 1906 cited from Dallai, 1979:275), the 'l' teeth (Massoud & Vannier, 1965 cited from Dallai, 1979:277), and with a smaller tooth (Bp) implanted about half way between the inner edge of the unguis and the posterior lamella (Dallai, 1975:275, Plate VIII,1,2). The outer region shows the two 'l' teeth united at the base to form a little slab from the centre of which the two lamellae rise (Dallai, 1975:275, Plate VIII,2). The hydrophile outer sclerite of the unguis is split basally to form an anterior and posterior laterobasal tooth. In Neelus murinus and Megalothorax minimus, the unguis bears at its basis an anterior and a posterior inner proximo-distal ridge. The ridges are diverging from the ungual body, but connected to it by a thin sheat and form as such kind of pseudonichia. The large anterior and posterior pseudonichium-like tooth of Neelus murinus and Megalothorax minimus are not homologue to the 'l' teeth of Neelides folsomi: while the former are part of the anterior and posterior facets of the unguis, the latter are part of the outer facet (Dallai, 1979:Plate XI,3;Plate XII,2;Plate VIII,1,2). This is confirmed in Megalothorax minimus, in which both the laterobasal 'l' teeth as well as the pseudonichium-like teeth of the anterior/posterior facet are present (Dallai, 1979:Plate XII,2).

Origin of the collembolan unguis

Known for being so concise and methodical (Miall, 1895:363), de Geer (1743) describes painstakingly and trustworthy the claws of Podura fusca and compares the collembolan claws with that of a lobster:

Fig.1. From subchelate foot to unguiate foot with reduced dactylus.
Frans Janssens © 2004.

Inspired by Lawrence (1999), I presume that the collembolan foot is derived from the ancestral precrustacean propodite-dactylopodite in a morphogenic sequence of subsequent laterally produced propodite, asymmetrically extended with lamellae, and laterally translocated and reduced dactylopodite (fig.1). In such a sequence we can distinguish the following types of limbs:

- subchelate limb:

the ancestral unmodified distal limb end, comprising propodite and terminal dactylopodite;

- semichelate limb:

a relatively small monolateral production is formed at the outer distal end of the propodite while the dactylopodite is dislocated laterally in opposite direction; the production constrains the articulation of the dactylopodite, and as such a primitive clasping device is formed

- unguichelate limb:

the semichelate propodite production is apicolaterally extended with lamellae, comprising the unguis, that eventually is as long as the dactylopodite, forming as such a more effective pincer-like grasping device

- unguiate limb:

the dactylopodite is reduced up to such an extent that the unguis is much longer than the dactylopodite; the condyles of the dactylopodite are reduced and the intrinsic muscles of the propodite that operate the dactylopodite are joined into one tendon, making the movement of the dactylopodite passive; the active grasping function of the foot is lost.

- unguiate limb with vestigial dactylopodite:

the dactylopodite is reduced into an immobile integumental tubercle, the so-called unguicular tubercle; this foot form is typically found in the poduromorph Hypogastruridae: Paraxenylla, Xenylla, Pseudacherontides, Acherontiella, and Schoetella, Odontellidae: Superodontella, and Odontellina, Brachystomellidae: Brachystomella, Neanuridae: Friesea, Pseudachorutes, Pratanurida, Protachorutes, Rusekella, Pseudachorudina, Cassagnaudina, Anuridella, Anurida, and Gastranurida, Tullbergiidae: Neonaphura, and Poduridae: Podura.

- unguiate limb with unguiculus:

in many genera, the unguicular tubercle developed secondarily at its anterior side a lamellate outgrowth, the unguiculus; this is the most common foot form in collembolans.

To conclude: the collembolan foot is a modified semichelate limb end of which the short 'fixed finger' of the propodite has been extended apicolaterally with lamellae to form the unguis and of which the 'moveable finger' has been strongly reduced into the unguicular tubercle, with optional secondarily developed unguiculus.

Dominant foot position

Fig.kw. Foot position in Dicyrtomina.
Valentine, B. © 2008.01.18.

Christiansen (1966:531-) observed the walking behaviour of specimens of entomobryid species in experimental conditions, and concluded that the most common foot position applied to a horizontal, firm, wet substrate is with the outer ungual face making contact with the substrate. Assuming such a wet substrate is the most common one in soil inhabitant collembolan habitats and observing the rich variety in outer ungual face features among the collembolan species, one can generalise that this foot position is the most dominant one for any Collembola species. In this foot position we can compare the function of the outer side of the ungual lamellae with the function of the sole of the human foot, and we might compare this way of walking of collembolans with the knuckle-walking of gorilla's (see fig.kw).

Fig.fp. Foot position.
Frans Janssens © 2003.

In fig. fpA, the unguis is flexed entad by the ungual flexors. This so-called dominant foot position is often found in specimens that are preserved in ethanol. The ungual flexor muscles shorten due to the action of the preservative and will flex the unguis entad. Note that in this foot position, the unguiculus may serve as the movable finger of a pincer. It is apressed towards the unguis whenever the flexor muscles shortens. Taken into account that the unguiculus is bilaminate at the side directed to the unguis in all but poduromorph collembolans, and that the unguis itself, often has an inner, longitudinal, median lamella, this pincerlike construct may serve as an effective clasping mechanism. Another function of the unguiculus is to restrict the entad flexing of the unguis while walking. In fig. fpB the unguis is extended in-line with the proximodistal axis of the limb. This foot position is often found in specimens that are placed in a clearing medium for examination with the microscope. The medium rehydrates and relaxes the ungual flexor muscles and due to the elasticity of the intugement the unguis will straigthen itself.

Pseudotarsus

Fig. bsock. Isotoma caerulea
ball and socket monocondyle of pseudotarsus
outer aspect
Frans Janssens © 2004.

Fig. b+s. Schema of ball and socket monocondyle of pseudotarsus
lateral section
Frans Janssens © 2004.

The pseudotarsus articulates with the tibia through a ball-and-socket joint (fig. bsock) (the ball making part of the tibia, the socket making part of the pseudotarsus) at the ectad base side of the unguis, such as in the tomocerid Tomocerus minor and in the isotomid Isotoma caerulea, and with an elastic articular membrane circumflexing the tibial apex. The ball is formed by an integumental bulging of the tibial (Tib) apical infolded cuticle that corresponds with an impression of the opposite pseudotarsal (PsT) wall (the socket, Sck) (fig. b+s). This construct allows the pseudotarsus to hinge inside and outside the tibial apex (note the syndetic articular membrane, am).

Fig. psex1. Schematic reconstruction of pseudotarsus in extended position achieved by haemolymph hydraulics Frans Janssens © 2004.

Fig. psex2. Schematic reconstruction of pseudotarsus in extended position achieved by intugement stiffness Frans Janssens © 2004.

Fig. psfl. Schematic reconstruction of pseudotarsus in flexed position achieved by flexor muscles Frans Janssens © 2004.

Three main pseudotarsal positions can be distinguished:

1. Due to increased hydraulic haemolymph pressure, the pseudotarsus is extended in its outer most position. High haemolymph pressures occur just before activating the furca. In this way the ungues are extended maximally to make sure that the ungues are released from the substrate surface and that the jump can be made effectively. The articular membrane is unfolded and stretched completely outside the tibia (fig. psex1). In this position, the unguicular tubercle is more or less lined up with the proximodistal axis of the limb.
2. A second extented position is achieved when the haemolymph pressure is reduced. Due to the elasticity of the intugement, and the stiffness of the syndetic ball-and-socket joint, the pseudotarsus takes a minimal extended position. The articular membrane is hinged and folded inside the tibial apex (fig. psex2).

3. Fig. flex. Isotomurus prasinus from Belgium
   Left metapseudotarsus in flexed position, outer aspect
   Frans Janssens © 2004.11.10

   The pseudotarsus hinges 'inside' the distal apex of the tibia (fig. psfl: fig. flex), up to such extent that it blocks against the inner pseudotarsal thickened ridge (see fig.u middle left; note that the pseudotarsus is tilted inside the transparent tibia); the articular membrane is unfolded and stretched completely, but now inside the tibia. Not only is the pseudotarsus hinged insed the tibia, the proximal inad wall of the pseudotarsus is unfolded (note that the fold closes in the extended position of the pseudotarsus). Note that the ridge is the modified distal edge of the ancestral tibia. This construct also serves as a shock breaking device. When the foot is placed in contact with the substrate, the circumflexing elastic membrane will stretch and slow down as such the hinging movement of the pseudotarsus. The final position of the hinged foot is sensed by the pseudotarsal setae, who as such serve as pseudotarsal position sensors.

Fig. f. Foldings in the articular membrane of the pseudotarsus
Frans Janssens © 2004.

In Orchesella flavescens the articular membrane of the pseudotarsus is folded accordeon-wise when the pseudotarsus is in its minimal extended position. The articular membrane folds are longitudinal as well as transvers (fig. f: Orchesella flavescens, left mesofoot, outer aspect, phasecontrast oc.10x obj.100x, immersion oil, negative image).

Additional apical tibial subdivisions, pseudotarsi, can be found in isotomids, such as a partially developed second pseudotarsus in Folsomia, Archisotoma (cf. Palissa, 1964:191), and Proisotoma (cf. Palissa, 1964:167), in the entomobryoid Microfalcula (cf. Betsch & Massoud, 1968:908), and in the tomocerid Tomocerus (cf. Palissa, 1964:232). A fully developed second pseudotarsus is found in all actaletids, such as in Actaletes (cf. Palissa, 1964:167) and Spinactaletes (cf. Soto-Adames, 1988:168), and in several tomocerids, such as in Pogonognathellus. In Folsomia, a transvers integumental fold at the apical inner side of the tibia marks the upper limit of a second pseudotarsus that has only partly developed. The fold does not yet circumflex the tibia, but ends midlaterally. The fold is the result of an integumental flexure. The tibial flexure acts as the second stage in a two-stage flexing mechanism of the unguis. It also serves as a shock damping system of the foot. When the body weigth is applied to the foot, the forces on the inner side of the tibia create a momentum on the inner wall of the tibia. As a result of the momentum, the tibia above the flexure is expanded, while the tibia below the flexure is compressed. In this way the fold opens at maximum. The lower part of the tibia gets hinged, semi-telescopically, into the open fold. When the body weight is released from the foot, the flexure unfolds and the tibia restores to its orginal condition due to the stiffness of the tibial intugement.
Pseudotarsogeny of the distal podomere is an ongoing evolutionary process in Collembola. It is not homolog to the subtarsogeny in Insecta.

Pseudotarsal tendon

Fig. ut. Series of micrographs (600x) at different levels of focus indicating the existence of a structure, that is an alternative to the unguitractor of Insecta, in Orchesella cincta, specimen from Belgium, ??? 2006.01.27, leg. Agnes ....
Frans Janssens © 2006.

Fig. ten. Pseudotarsal tendon attachement in Isotomurus sp., specimen from Belgium, ??? 2006.??.??, leg. Agnes ....
Frans Janssens © 2006.

In the extended position of the unguis (U), the pseudotarsus is hinged extad due to haemolymph hydraulic pressure, rotating at the outer monocondyle (C1). The tendon (Te) runs upwards into the tibia (Ti) (fig.ten; Isotomurus sp., metafoot, oc.10x obj.60x). The connection point of the tendon (Te, see arrow) to the pseudotarsal body wall is located at the upper inner side of the pseudotarsus. Notice a second smaller monocondyle (C2) at the basis of the unguiculus (u).

Three parts can be distinguished that play a functional role:
1. the basis of the pseudotarsal tendon
2. the entad pseudotarsal body wall
3. the unguicular tubercle

Fig. t. Pseudotarsal tendon with delta shaped connecting fibers.
Frans Janssens © 2004.

Pseudotarsal tendon basis. While the pseudotarsal tendon itself is lintshaped, distally, just before it is attached to the pseudotarsus, the tendon is medially and laterally extended with fibers forming as such a delta shaped tendal basis (fig. t: Orchesella flavescens, left mesofoot, outer aspect, phasecontrast oc.10x obj.100x, immersion oil, negative image).

Fig. d. Schematic view of pseudotarsal tendon connecting fibers in cross-section.
Frans Janssens © 2005.

The delta shaped connecting fibers are organised into 3 sets: the median set of fibers is fused with the internal side of the articular pseudotarsal membrane, and runs further down to the base of the unguicular tubercle, while the two lateral sets of fibers are fused with the external side of the articular pseudotarsal membrane, running down to the upper rim of the pseudotarsal body wall. This construct makes the fibers in cross-section U-shaped with backfolded edges (Fig.d: folded pseudotarsal articular membrane: green; tendal fibers: red). The concave side of the U-shape is directed towards the entad extruded pseudotarsal body wall. The lateral fibers connect the tendon with the pseudotarsal articular membrane by encompassing the pseudotarsal basal open end (see the lateral views on these connective fibers in the schematic figs psex1, psex2, psfl).

Entad pseudotarsal body wall. The median set of tendal connecting fibers is not only fused with the internal side of the articular pseudotarsal membrane, but runs further down to the base of the unguicular tubercle and is fused with the internal side of pseudotarsal body wall. This median set of fibers runs to the unguicular tubercle via an extruded section of the entad pseudotarsal wall, as seen typically in Entomobrya multifasciata and Isotoma caerulea.

Unguicular tubercle. The median set of tendal connecting fibers terminate at the unguicular tubercle, which serves as anchoring point of the pseudotarsal tendon.

Muscles that operate the unguis

Fig. sect. Pseudotarsal tendon section in Podura aquatica., from Belgium, ??? 2006.??.??, leg. Janssens, F.
Frans Janssens © 2006.

tendon >

In many Collembola, such as Archisotoma and Pogonognathellus, a long apodeme arises from the proximal base of the unguiculus, thus opposite to the unguis, where the unguicular tubercle is located. This apodeme, halfway the tibia, is inserted with two sets of two muscular fibers. One set of fibers connects to the outer proximal base of the tibia (tibial fibers). The other set passes through the tibiofemoral joint into the femur (femoral fibers).
The tibia is a hollow cuticular tube that is internally lined with a layer of epidermal cells (fig. sect. cross section of the tibia of a specimen of Podura aquatica from Belgium, by Nico Büscher). The pseudotarsal tendon or apodeme is located in the center of the tibia. The tendon itself has a lint shaped section (amber colour). Close near to the tendon, the section through the tibial fibers of the pseudotarsal flexor or retractor muscle can be observed (red colour).

Comparison of walking limbs of Crustacea, Collembola and Insecta

• Stage 1: the most primitive ancestral stage of the arthropodium: coxopodite (fig. x 1:D) with telopodite composed of two podomeres, the basal meropodite and distal propodite (fig. x. 1:C,B), and a clawlike terminal podomere, the dactylopodite (fig x. 1:A).
• Stage 2: the basal podomere of the telopodite, the meropodite, is differentiated into two subpodomeres, the basal prae-ischium and distal prae-merus (fig. x 2:C1,C2). Associated condyles and intrinsic muscles have been formed. Note that in this stage the basal subpodomere, the prae-ischium, (fig. x 2:C2) contains two pair of antagonistic muscle sets.
• Stage 3: the original muscles of the basal podomere of the telopodite, the meropodite, are shortened and retracted in the distal subpodomere, the merus (fig. x 3).
• Stage 4: the distal podomere of the telopodite, the propodite, is differentiated into two subpodomeres, the basal prae-carpus and distal prae-propodus (fig. x 4:B1,B2).
• Stage 5: the groundplan of the crustacean walking limb: basis, ischium, merus, carpus, propodus, and dactylus.

Fig.x. Simplified scheme of the morphogeny of the walking limb of
Collembola and Insecta in relation to there crustacean ancestors.
Frans Janssens © 2004.

Distal telopodomeres. Grimaldi (2001:1158) considers the tibiotarsus of Collembola as a compound podomere, being homolog with the fused hexapodan tibia and tarsus (tarsi). However, more parsimonous, we presume that the penultimate podomere of Collembola, conventionally called tibiotarsus, is homolog with the propodite of the stage 2 crustacean endopodite. Therefore, it is not a compound podomere. It is homolog with the tibia of Insecta.
Note: the scheme of splitting telopodomeres can be further recursively extended. Each ancestral distal telopodomere can be split in two apomorphic podomeres in two stages as described above. To comply with the groundplan of the Paleozoic arthropodan telopodite with 7 telopodomeres as defined by Kukalova-Peck (2008): prefemur, femur, patella, tibia, basitarsus, eutarsus and pretarsus, 4 additional stages 6 to 9 in the scheme are required. The ancestral insect leg can then be derived from stage 9.
In other words, Collembola branched from the crustacean line before Insecta did and independant from them. This model is in line with the findings of Spears & Abele (1997, cited from Lange & Schram, 1999) and the molecular studies of Shao, Zhang, Ke, Yue & Yin (2000) and Nardi et al. (2001).

Terminal telopodomere. The terminal walking limb podomeres of Collembola and Insecta are difficult to compare, morphologically. They have developed from the terminal podomere of the ancestral telopodite, the dactylopodite, independant from each other. In Collembola, the terminal clawlike podomere reduced into a redundant process, the unguicular tubercle, while a secondary claw, the unguis, developed opposed to it. The apodemes of the flexor and extensor muscles of the reduced claw fused distally into a long common tendon (not shown in fig. x.). The head of the original flexor muscles transposed to the femoral base, while the head of the original extensor muscles remained attached proximally to the tibia. Both muscle sets function synchroneously as flexors of the new unguis. The primary condyles of the original claw were reduced as well, while a secondary monocondyle was formed to support flexing the new unguis (Fig. x 2a: secondary condyle in green), giving rise to a new terminal limb subdivision. Note that this subdivision is not homolog with the tarsus of Insecta. In this paper it is called 'pseudotarsus' to distinguish it from the true tarsus of Insecta (the tarsus is derived from the dactylopodite, while the pseudotarsus is derived from the propodite). Secondarily, the unguicular tubercle might optionally be extended with lamellae to form the unguiculus. In Insecta, the terminal clawlike podomere itself was subdivided into several subdivisions lacking intrinsic muscles. At the terminal subdivision two parallel ungues were formed. The original flexor muscle set split in two muscle sets: one remained flexor of the tarsus, while the other developed a long tendon and served as flexor of the ungues. The head of the new ungual flexor muscles split as well and one part was moved into the femur (fig. x 3a). Note that the unguis and ungual flexor of Insecta are not homolog to the unguis and ungual flexor of Collembola: the unguis of Insecta is derived from the dactylopodite, while the unguis of Collembola is derived from the propodite; the ungual flexor of Insecta is derived from the flexor of the dactylopodite, while the ungual flexors of Collembola are derived from both extensor and flexor of the dactylopodite.

Revised distal podomere terminology

Within the context of our hypothesis, the concept of 'tibiotarsus', as the name suggests, a compound podomere comprising the tibia and tarsus, should be revised: 1. it is not a compound podomere, but a direct derivate of the ancestral propodite (no fusion of podomeres is involved), 2. it includes the unguis, it being a feature of the distal part of the tibia, the pseudotarsus, a secondary derived subpodomere. The 'pretarsus' requires redefinition that refers only to the unguiculus.

Fig.2a. Architecture of foot complex of Tomocerus minor.
Photo Keith Brocklehurst © 1999.

Applying this model to the foot of Tomocerus minor (fig. 2a) may serve as an illustration of the ancestral architecture of the collembolan terminal podomeres. The tibia (Ti) bears distally a flexible subpodomere - conventionally called 'pretarsus', also called 'post-tarsus' (Bitsch & Bitsch, 2000:140) expressing effectively that it is not part of the tarsus - that is in this context called 'pseudotarsus' (PsTa). The tibia and pseudotarsus are extended, in-line with the proximodistal axis (Pda) of the limb, with the unguicular tubercle, (optionally) extended with the unguiculus (ung) - also called 'empodium' or 'empodial appendage'. Distally, the pseudotarsus bears laterally at the outer side a large hollow integumentary process, the pseudotarsal spine (Sp). The pseudotarsal spine is typically latero-distally extended with lamellae (Lam). The ectad lateral pseudotarsal spine bearing extended lamellae is called the 'unguis' (Ung) - conventionally called 'pretarsal claw', but is actually the 'pseudotarsal claw'.
Note: this presumption is an alternative to Sobotnik & Stys (1998) who suggest that the median claw of the Parainsecta (Collembola and Protura) may 'have evolved by transformation of [the] post-tarsus with lateral claws already present' (cited from Bitsch & Bitsch, 2000:140).

Taking into account that the foot of Tomocerus minor displays several derived features, the following references of figures of the feet of representative unguiculus bearing poduromorphs may serve as some more illustrations of the ancestral crustacean architecture of the distal podomeres as described above:
Ceratophysella gibbosa (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:120,fig.54D)
Hymenaphorura hispanica (Onychiuridae) (Jordana et al. in Ramos et al., 1997:637,fig.240D)
Hypogastrura sahlbergi (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:89,fig.37F-G)
Kalaphorura burmeisteri (Onychiuridae) (Jordana et al. in Ramos et al., 1997:626,fig.236B)
Metaphorura denisi (Tullbergiidae) (Simón & Luciáñez in Ramos et al., 1997:701,fig.263C)
Microgastrura sensilata (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:186,fig.85B)
Mucrella acuminata (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:129,fig.59C)
Onychiurus rectospinatus (Onychiuridae) (Jordana et al. in Ramos et al., 1997:534,fig.207B)
Protaphorura campata (Onychiuridae) (Jordana et al. in Ramos et al., 1997:569,fig.217B)
Schaefferia emucronata (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:202,fig.91C)
Triacanthella perfecta (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:232,fig.105D)
Typhlogastrura mendizabali (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:181,fig.83E)
Willemia denisi (Hypogastruridae) (Jordana & Arbea in Ramos et al., 1997:213,fig.96C)

Foot with functionally 'rolling' unguis

Fig.u. Simplified reconstruction of walking foot showing the 'rolling' unguis.
Frans Janssens © 2004.
From left to right:
a. posterior aspect of foot with unloaded and not flexed unguis;
b. posterior aspect of foot with unloaded but flexed unguis;
c. posterior aspect of foot with springloaded flexed unguis, unguis bending due to body weight;
d. posterior aspect of foot with springloaded flexed unguis, unguis fully bent due to body weight.

Just before putting the foot on the smooth or wetted substrate, the ungual flexors pull the pseudotarsus upwards, and it hinges inside the tibial apex. When the foot touches the substrate surface, the unguis will start bending entad due to the body weight applied to the feet. The unguis will continue to bend up to such extend that the ungual lamellae (lateral (reddish) and inner (yellowish) longitudinal lamellae) are fully bent together with the tip of the ungual body itself (see fig.u. middle and outer right). Now, the unguis is springloaded and the elastic energy is released when the limb is lifted, reducing in this way effectively the required effort to be spent by the limb levator muscles. Note that the bending unguis functionally behaves as a wheel: it actually rolls over the substrate surface.
Note that many species have lateral teeth at the edges of the lateral longitudinal ungual lamellae (see fig.7). These teeth will penetrate the surface texture while the unguis 'rolls' over the substrate. The lateral teeth improve the stability of the foot position in contact with the substrate.
The onychiurid Ongulonychiurus colpus Thibaud & Massoud, 1986 may serve as an example of extreme adaptation to this rolling unguis condition. This cave species has very long and slender ungues, with a length equal to that of tibia and femur together.

Conclusion

Acknowledgements

Glossary

Literature

• Wolf, H. & Harzsch, S. 2002. Evolution of the arthropod neuromuscular system. 1. Arrangement of muscles and innervation in the walking legs of a scorpion: Vaejovis spinigerus (Wood, 1863) Vaejovidae, Scorpiones, Arachnida., Arthropod Structure & Development, xx, 2002, p.yy-zz.