http://www.collembola.org/publicat/thorax1.htm - Last updated on 2012.11.28 by Frans Janssens
Checklist of the Collembola: Note on the Morphology of the Prothorax

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

Introduction

The sister group of the Poduromorpha, the 'higher Collembola', here called the 'new Collembola', i.e. Neelipleona + Entomobryomorpha s.l. + Symphypleona, is characterised by the reduction of the prothorax (Moen & Ellis, 1984:203). This argument has not received wide acceptance (Hopkin, 1997:28).
In this paper, we will show that the reduction of the prothorax is particulary limited to the protergum and that this protergal reduction not only is a characteristic of the so-called 'new Collembola', but to all Collembola. Based on an analysis of the tergal labioprothoracic chaetotaxy and the labioprothoracic muscular system it is shown that the protergum must have been reduced already in the common collembolan ancestor.
The protergal reduction or the apparent loss of the protergite may have resulted from the fusion of part of the dorsal area of the prothorax to the posterior part of the head capsule or to the anterior part of the mesotergum, so forming a large composite pro-mesotergal sclerite. This latter view is supported by the opinion of some authors that claim that some motor muscles of the procoxae are dorsally attached to the anterior part of the so called 'mesonotal' plate (Carpentier, 1947, 1949; Bretfeld, 1963; François, 1996 cited from Bitsch & Bitsch, 2000:138-139).

"Ici, la plupart des muscles longitudinaux dorsaux de la tête ne sont pas relayés par des thoraciques, en raison de la réduction du prothorax. Une seule paire de muscles céphaliques passe au thorax et gagne le mésonotum sans s'être attachée au foramen. Tout sela prouve que la tête globuleuse est secondaire."
(Denis in Grassé, 1949:121 concerning the cervix ("constriction collaire") of Anurida).
Contrary to Carpentier (1949), who suggested a promesotergal fusion in Entomobryomorpha, and more in line with Denis (1949), who presumed that the globulose head capsule is a derived character, we hypothise that the protocollembolan protergite was subdivided transversally in a broad anteroprotergite and a narrow posteroprotergite. The anteroprotergite fused with the posterior segment of the head capsule. The remaining posteroprotergite lost its function as location of muscular attachement in the 'new Collembola', reduced further, and eventually also fused with the head capsule.

Reduction of the protergite

We hypothise that the reduction of the protergite, that is generally considered as an ordinal characteristic of the Entomobryomorpha, is a more general characteristic of all Collembola. What is commonly described as the protergite of recent Collembola is in fact only the posterior part of the ancestral protergite. We believe that the ancestral protergite has been subdivided into a relative broad anterior part and a narrower posterior part. In the protocollembolan, the anterior part already fused with the labial tergite. In other words, the protergal reduction is a characteristic that is common to all Collembola. The posterior part reduced further stepwise in the ordinal phylogeny. With as initial step a not reduced posterior part in the Poduromorpha. And with a further reduction as final step in the "Neocollembola": Neelipleona, Entomobryomorpha s.l. and Symphypleona.

Fig.r. Rhyniella praecursor
After Scourfield (1940).
A fossil perspective
Certainly, the reduction of the protergite has occured by the Devonian (400 million years ago), since Rhyniella praecursor Hirst & Maulik, 1926, the oldest known collembolan fossil from the red sandstone Rhynie chert beds, Scotland, according to Scourfield (1940) and confirmed by Greenslade & Whalley (1986:320), lacks a protergite. Greenslade & Whalley (1986:321) classify Rhyniella praecursor within the family Isotomidae. This implies that the protergal reduction must have been initiated in a rhyniellian ancestor older than Devonian age. In all known Collembola where the protergite is still present, such as Poduromorpha, it is always much smaller and it has less transversal rows of setae than the other thoracic tergites.

An embryological perspective
Poduromorphan embryology: Claypole (1898) conducted a detailed study on the embryonic development of Anurida maritma. Claypole (1898:278) concludes that Anurida shows characters allying it with crustaceans and myriapods rather than insects . Anurida is compared to a sexually matured insect embryo. As a form that has lengthened its embryonic life and has shortened its adult life. Unfortunately, nothing is said about the prothoracic development.
Entomobryomorphan embryology: Uemiya & Ando (1987,1991) conducted a detailed study on the embryonic development of Tomocerus ishibashii. The embryonic development can be divided into nine stages. During the sixth stage, the dorsal region of the prothoracic segment begins to develop at a slower rate than the mesothoracic and metathoracic segments. The protergum becomes narrower and much less sclerotised, one of the characteristic features of Entomobryomorpha s.l. (Hopkin, 1997:144-145).

A chaetotactic perspective
In accordance with the regression of the prothorax, the prothoracic chaetotaxy has been simplified (Cassagnau, 1974:302-303).
The dorsal chaetotaxy of the thoracic segments basically consists of three rows of setae, 'a', 'm' and 'p', as proposed by Yosii (1956) (André, 1988:514). The naming of the last two cephalic rows of setae [in Xenylla] of da Gama, namely cervicals (c) and parietals (p) is not generally followed. Yosii (1956) proposed to name the most posterior row of the dorsal cephalic setae, 'c', and the row immediately before, 'p'. In 1960, Yosii changed his mind and reversed the two namings. da Gama [(1980)] still follows the former terminology while Cassagnau (1975) and others have adopted the latter (André, 1988:508). Thibaud (1970:134,142,144) in Ceratophysella bengtssoni (fig.10A,B), in Schaefferia coeca (fig.11A,B), and in Typhlogastrura balazuci (fig.12A,B) describes the posterior region of the head as c-row and p-row setae with respectively 4 and 7 setae, and refers to the protergal setae as the medial row. Cassagnau (1974:301[fig.1],302,303) describes in his model of the cephalic chaetotaxy of the Poduromorpha the occipital region of the head as c-row and p-row setae, with respectively 4 and 5 setae, and refers to the protergal setae as the anterior row or posterior row, which he considers more evolutionary stable than the medial row (1974:304). André (1988:509[fig.2B],514[fig.6A]) describes in the phanerotaxy of Xenylla jocquei the posterior region of the head as c-row and p-row setae and refers to the protergal setae as the anterior row. Jordana & Arbea (1997:62[fig.23]) in Hypogastruridae follow the original terminology of c-row and p-row of setae at the posterior region of the head and refer to the protergal setae as the medial row, which is in-line with Yosii (Cassagnau 1974:304). Fjellberg (1998:15,30[fig.10H]) describes in the chaetotaxy of the Poduromorpha the posterior region of the head as c-row and p-row setae and refers to the protergal setae as the posterior row.
Fig.cc. Nomenclature of cephalic setae in basal Hypogastruridae (Modified after Cassagnau (1974))
Janssens, F. 2004.

We define an alternative schema for the setal pattern on these regions that is derived from Cassagnau (1974:301) but takes into account the cephalic chaetotactic model of Thibaud (1970) and the protergal chaetotactic model of Fjellberg (1998). In our postoccipito-protergal chaetotactic model, the poduromorphan cephalic 'cervical' and 'parietal' rows of setae and the prothoracic posterior row of setae are derived from an ancestral protergite with generalised anterior, medial and posterior row of setae.
A transvers region of desclerotisation of the protergite inbetween the medial setae and posterior setae subdivided the protergite into two sclerites: the anteroprotergite and the posteroprotergite. This secondary dorsolateral membraneous region improved the free dorsoventral movement of the head capsule relative to the trunk.
The broad anteroprotergite bears the anterior and medial row of setae. The anteroprotergite fuses with the tergite of the posterior cephalic segment. The ancestral prothoracic anterior row of setae (a-setae) becomes the by Yosii (1960) designated cephalic row of cervicals (c-setae). The ancestral prothoracic medial row of setae (m-setae) becomes the by Yosii (1960) designated cephalic row of parietals (p-setae).
The narrow posteroprotergite bears the remaining ancestral posterior row of setae (p-setae), in-line with Fjellberg's protergal model (1998).
In the hypothetical primitive homochaetose scheme of Cassagnau (1974:305), each thoracic demi-tergite counts 7 setae in the a-row and m-row and 6 setae in the p-row; setae m7 and p4 are sensilla. The cephalic c-row and p-row count only 5 setae. In our scheme (fig.cc), we redesignate several setae of the c-, p-, g- and l-rows, in-line with Thibaud (1970) and to reflect our labioprotergal fusion hypothesis, as such:
setae g1 and g1' of Cassagnau (1974:301) respectively become p6 and p7 of Thibaud (1970,134,142,144);
setae l0 and l1 respectively become c6 and c7; and
seta g2 replaces c5 and vice versa.
In this way, the cephalic p- and c-row have each a complete row of 7 setae, complying to the ancestral scheme as proposed by Cassagnau (1974:305). The g-row is reduced to g2-5; the l-row is reduced to l2-l5.

Open issues:
What with seta p7 as sensillum?
In Onychiuridae: the posteroprotergite appears to bear more than just the p-setae. Is this a secondarely derived character, the result of polychaetosis? Or are the Onychiuridae the most archaic poduromorphans, and is the primitive chloting polychaetose and not homochaetose? Are Onychiuridae an early side branch with a different, larger posteroprotergite? And/or desclerotisation between a-row and m-row in stead of between m-row and p-row?

Fig.m. Schematic musculature of three subsequent segments
Janssens, F. 2002.
A muscular perspective
Basically, any arthropodan tergum comprises a narrow anterior acrotergite, which posterior border is marked by a transvers invagination, the antecosta, which is characterised externally by an antecostal suture, followed by the sclerotised hardened eutergite and the non-sclerotised membraneous conjuctiva.

In a generalised and simplified schema (fig.m: threedimensional caudolateral reconstruction of a longitudinal median section through the thoracic segments) of the ancestral arthropodan thoracic musculature, the following two basic categories of sets of muscles can be distinguished:

  1. intrasegmental muscles (blue muscle fibers), comprising
    1. anterior lateral transverse tergosternal muscles,
    2. posterior lateral transverse tergosternal muscles,
    3. medial oblique transverse tergosternal muscles;
  2. intersegmental muscles (yellow muscle fibers), interconnecting the antecostae of subsequent segments, comprising
    1. longitudinal tergal muscles,
    2. longitudinal sternal muscles,
    3. oblique transverse tergosternal muscles.

Fig.m2. Schematic musculature
of prothorax and mesothorax
Janssens, F. 2002.
Of interest in our hypothesis of the ancestral arthropodan prothoracic musculature (fig.m2: threedimensional lateral reconstruction of prothoracic and mesothoracic musculature) are the following four basic sets of muscles:
  1. the intersegmental longitudinal protergal muscles (yellow) (dorsal set), the anterior end mediodorsally attached to the protergal antecosta and the posterior end mediodorsally attached to the mesotergal antecosta;
  2. the intrasegmental medial dorsoventral protergosternal muscles (magenta), one end mediodorsally attached to the middle of the protergite and the other end medioventrally attached to the prosternal apophyses/antecosta;
  3. the intrasegmental lateral dorsoventral protergosternal muscles (blue) (anterior and posterior set), one end laterodorsally attached to the protergite and the other end lateroventrally attached to the prosternite;
  4. the oblique promeso-intersegmental muscles (green), the anterior end medioventrally attached to the prosternal apophyses and the posterior end mediodorsally attached to the mesotergal antecosta.

(Denis in Grassé, 1949:121) concerning the muscles of Anurida:
"Muscles tergo-tentoriaux. Les uns viennent du thorax et sont souvent intersegmentaires avec deux innervations distinctes."
These muscles partly map

(Denis in Grassé, 1949:123) concerning the muscles of Anurida:
"Muscles longitudinaux dorsaux. Il y a des longitudinaux dorsaux dans la région postérieure de la tąte; ils appartiennent assurément au segment labial, mais il est impossible d'affirmer qu'ils n'appartiennent qu'à lui, recevant leur innervation du nerf r (fig.16) certainement complexe. Le bord postérieur de la capsule céphalique répond è la limite thoraco-labiale et c'est à son niveau que les longitudinaux dorsaux, céphaliques et thoraciques, portent leurs plaques motrices. Celles-ci résultent de la fusion des plaques motrices, destinées aux muscles céphaliques, des nerfs r avec celles des nerfs intercalaires (nerfs prothoraciques, tout au moins en partie) destinées aux muscles thoraciques. Cette jonction des nerfs atteste la condensation des domaines mitoyens à deux tagmes."
These muscles map to the longitudinal protergal muscles of the generalised schema, of which the anterior end is attached to the protergal antecosta which is fused with the labial tergite and of which the posterior end is attached to the mesotergal antecosta.

The generalised prothoracic muscular model can be mapped onto the muscular system of Orchesella and Neanura as described by Bretfeld (1963).

The following cephaloprothoracic muscles in Neanura (Schaller, 1970:30[fig.45b]) correspond subsequentially with the original prothoracic basic muscles:

Fig.y. Neanura sp.
Thoracic musculature
After Schaller, F. 1970.

1. the longitudinal cephaloprotergal muscles (Lbdlm1+Idlm1,Lbdlm2+Idlm2), one end mediodorsally attached at the head capsule and the other end mediodorsally attached at the mesotergal antecosta;
2. the dorsoventral cephaloprosternal muscles (Lbldvm1), one end medioventrally attached at the prosternal antecosta and the other end mediodorsally attached at the middle of the headcapsule.
3. the lateral cephaloprosternal muscles (Lbldvm2a,b), one end laterodorsally attached to the head capsule and the other end lateroventrally attached to the prosternite;
4. the promeso-intersegmental muscles (Iism1b), one end medioventrally attached at the prosternal antecosta and the other end mediodorsally attached at the mesotergal antecosta.

It is necessary to carefully reinterprete Bretfeld's schema of the thoracic musculature of Neanura (Schaller, 1970:30[fig.45b]): comparing the muscles of the labial segment, the prothorax and the mesothorax shows the following analogy:
- the intersegmental muscles ism1-2 of the mesothorax map to the muscle ism1b of the prothorax
- the dorsoventral muscles dvm2 of the mesothorax map to the muscle ldvm1 of the labial segment
- the lateral dorsoventral muscles ldvm1-4 of the mesothorax map to the muscles dvm1-3 of the prothorax
The protergal antecosta that is shown in Bretfeld's schema is in fact an endoskeletal ridge of the posteroprotergite to which are attached the lateral dorsoventral muscles of the prothorax.

The following cephaloprothoracic muscles in Orchesella (Schaller, 1970:30[fig.45a]) correspond subsequentially with the original prothoracic basic muscles (see fig.x):

Fig.x. Orchesella sp.
Thoracic musculature
After Schaller, F. 1970.

1. the longitudinal cephaloprotergal muscles (Idlm), the anterior end mediodorsally attached to the head capsule and the posterior end mediodorsally attached to the mesotergal antecosta; note that the muscles are not attached to the protergum;
2. the medial dorsoventral cephaloprosternal muscles (Lbism), one end mediodorsally attached to the middle of the posterior part of the head capsule and the other end medioventrally attached to the prosternal apophyses; note that the muscles appear to be intersegmental: from the labial segment to the prothorax;
3. the lateral cephaloprosternal muscles (Lbldvm), one end laterodorsally attached to the head capsule and the other end lateroventrally attached to the prosternite; note that the muscles appear to be intersegmental: from the labial segment to the prothorax;
4. the promeso-intersegmental muscles (Iism3&4), the anterior end medioventrally attached to the prosternal apophyses and the posterior end mediodorsally attached to the mesotergal antecosta.

Labioprotergal fusion in Poduromorpha

In all arthropods with well-developed body-wall sclerites, the definitive segmentation is a secondary one; but the limits of the secondary segments differ according to the relations of the sclerotisation in the primary intersegmental regions to that of the segmental regions beforeand behind them (Snodgrass, 1935).

Fig.XX. Schematic labioprothoracic segmentation.
A. Primary segmentation; B. Protergal desclerotisation; C. Labio-protergal fusion.
Modified after Snodgrass, R.E. 1935.

In the ancestral collembolan labioprothoracic segmentation, where both the dorsal and the ventral primary intersegmental areas of sclerotisation are continuous with the segmental sclerites following, the typical arthropodan secondary segmentation prevails, and the functional intersegmental rings are the membranous posterior parts of the primary segments (fig. XX A, Mb). The primitive intersegmental fold (Isg) forms an antecosta (Ac) with externally an antecostal suture on both the tergum (T) and the sternum (S), and the precostal lip forms an acrotergite (atg) of the tergum and an acrosternite (ast) of the sternum.
In a first phase, the desclerotisation of a transvers protergal region inbetween the medial row of setae and the posterior row of setae, and assuming that the ventral sclerites retain the form typical of secondary segmentation, the protergal sclerotisation takes the form of an independent intertergite and eutergite. The dorsal half of the labioprothoracic intersegmental membrane embraces the intertergite (fig. XX B) and includes a part of the two adjoining primary segmental regions. The intertergite can be seen as the tergal antecosta having precostally a narrow intertergal acrotergite and postcostally a broad intertergal posttergite (ptg) bearing the anterior and medial rows of setae.
In a second phase, the intertergite is united with the labial tergum preceding (fig. XX C). Since the intersternal sclerotisation remains continuous with the prosternum following, the labioprothoracic conjunctival membrane (Mb) crosses obliquely on the side of the body ventrally from the posterior part of the labial segment to the anterior part of the prothoracic segment dorsally, and the postcostal intertergal posttergite becomes the posttergite of the labial segment.

The protergite of the ancestral protocollembolan is subdivided (fig.hp) into an anteroprotergite (B) that fused with the labiotergite to become an integrated part of the head capsule and a posteroprotergite (A) that becomes the 'protergite' of the extant Collembola.

The dipluran 'micronotum', as described by Verhoeff (1904) (cited from Denis in Grassé, 1949:167), might be homolog with our hypothetical intertergite.
The proturan 'gnathotergite', which is an isolated tergite in front of their 'protergite' might also be homolog with the intertergite (the anterior sclerite of the ancestral protergite).
So it seems that the origin of the reduction of the protergite goes even further back in time. Assuming that the current phylogenetic views on the 'close' relationship between Collembola and Protura is correct, the process of protergite reduction seems to start even at the level of their common ancestor. While in Protura the anterior part of the ancestral protergite remained an independent sclerite and developed into the proturan gnathotergite, in Collembola it fused with the posterior segment of the head. The fact that proturan taxonomy uses a term as 'microcephalic' is also interesting to note: it emphasises the difference between the not fused protergite of the Protura (resulting in an apparently smaller head) and the fused protergite of the Collembola (resulting in an apparently larger head).

A morphometric perspective
Assume that the collembolan ancestor is derived from a hypotethical arthropod with homogeneous segments and generic locomotory appendages per segment. Assume that the anterior segments already fused to form the head capsule. While the abdominal appendages specialised functionally per segment, to form the furca of the fourth abdominal segment, the retinaculum of the third abdominal segment, and the ventral tubus of the first abdominal segment, the appendages of the second abdominal segment obliterated. As a result of the reduction of the appendages of the second abdominal segment the size of the sternum reduced proportionally. Due to the sternal reduction of the second abdominal segment, all anterior sterna are dislocated posteriorly. This can be observed best in Poduromorpha and Entomobryomorpha s.l.: in lateral view, the coxae appear to be intersegmental, ventrodorsally aligned with the tergal conjuctivae. Due to the posterior dislocation of the thoracic sterna, the head capsule becomes slightly tilted in the downward direction. Due to the downwards tilt of the head capsule, the protergum and propleura are exposed to additional stress. The additional stress in the cuticula is compensated by the labioprotergal fusion as described above.

Fig.hp. Hypogastrura sp.
Model for prothoraco-cephalic musculature.
Note on the position of the head
As discussed before, due to the posterior dislocation of the thoracic sterna, the collembolan head has an inherent tendency to be downwards directed. Muscles mainly involved in head positioning (Fig. hp) are : primary tergal longitudinal muscle from mesotergal antecosta (atII) to protergal antecosta (atI): 1, primary tergal longitudinal muscle from protergal antecosta to labiotergal antecosta (atl): 2, secondary tergal longitudinal muscle from mesotergal antecosta to labiotergal antecosta: 5, primary sternal longitudinal muscle from mesosternal antecosta (asII) to prosternal antecosta (asI): 3, primary sternal longitudinal muscle from prosternal antecosta to labiosternal antecosta (asl): 4, secondary sternal longitudinal muscle from mesosternal antecosta to labiosternal antecosta: 6, primary sternotergal tranversal muscle from mesotergal antecosta to prosternal antecosta: 7, primary sternotergal tranversal muscle from prosternal antecosta to mid protergum: 8 ("labial" tergosternal muscle), secondary sternotergal tranversal muscle from mesotergal antecosta to prolabial antecosta: 9 (intersegmental muscle).

NoFunctionNeanuraOrchesella
1Primary levatorI dlm1 + dlm2Idlm
2 Lb dlm1 + dlm2(absent)
3 I vlmIvlm
4Primary promotorLb vlm2Lbvlm2
5Secondary levator(absent)Idlm
6Secondary promotorLb vlm1Lbvlm1
7Primary promotorI ism1bIsm3u4
8Primary levatordvm2Lbldvm
9Secondary levatorI ism1aIism1 + Iism2
Muscles of Neanura and Orchesella after Schaller (1970).

Conclusion 1: Due to the increased longitudinal stress at the tergum, caused by the downwards tilted head capsule, the function of the primary tergal longitudinal muscle from mesotergal antecosta to protergal antecosta (no. 1) got enforced by replicating a secondary tergal longitudinal muscle from mesotergal antecosta to labiotergal antecosta (no. 5). In Neanura this secondary muscle is (still) lacking. In Orchesella it is fully developed.
Conclusion 2: Due to the labioprotergal fusion, the primary tergal longitudinal muscle from protergal antecosta to labiotergal antecosta (no. 2) lost its function, it became redundant and reduced. In Neanura it is still available. In Orchesella it is completely obliterated.
Both indicate that the Poduromorpha are phylogenetically more ancestral then the Entomobryomorpha.

A case study as 'proof of concept'

Fig.s. Synameria orientalis
With lateral tergal lobes
After Cassagnau, P. 1990.
The Himalayan neanurins Synameria miranda (Yosii, 1966) and Synameria orientalis Cassagnau, 1990 are characterised by a dorsoventral flattened body having prominent lateral tergal lobes ('lobes latéraux digitiformes') (see Fig.s). All thoracic terga as well as the anterior abdominal terga bear such lateral lobes. Remarkable is that such a lobe also occurs at the lateroposterior tergal region of the head.
The lateral region of the mesotergum and metatergum is kind of bilobed, having a prominent posterior lobe and a much less prominent anterior lobe. The protergum is not laterally bilobed. Both the cephalic lateral lobe and the not bilobed protergum are a kind of proof of concept for our hypothesis of the labioprotergal fusion: the cephalic lateral lobe corresponds with the anterior lobe of the primitive protergum; the protergal lateral lobe corresponds with the posterior lobe of the primitive protergum.

Chaetotactic confirmation, based on Synameria orientalis Cassagnau, 1990:21 Fig.1 :
1. the lateral macrosetae of the thoracic lobes: note the consistency in the number and position of the two lateral macrosetae on the lobe of the protergum and of the posterior lobe of meso- and metatergum.
2. the number of setulae of the posterior row of the thoracic terga: note the consistency in the number of setulae of the posterior row (p-row): 4 setulae.
Both are an indication that the posterior part of the meso- and metatergum is homolog with the protergum.
3. the number of setae of the mediodorsal region of the anterior row: the mediodorsal region of the anterior row (a-row) of the meso- and metatergum bears 4 setae; the mediodorsal region of the 'anterior row' of the posterior region of the head bears 4 setae.
4. the lateral cluster of anterior setae: the lateral region of the anterior lobe of the meso- and metatergum bears a cluster of three setulae and one macroseta; the cephalic lobe bears a cluster of three setulae and three macrosetae (a result of plurichaetosis?).
Both are an indication that the anterior part of the meso- and metatergum is homolog with the posterior tergal part of the head.

Hypertrophic evolution as 'proof of concept'

Fig.Lk. Leenurina khualaza
Protergite completely reduced
After Deharveng & al. 2011 Fig.1A

The neanurid genera Caputanurina Lee, 1983 and Leenurina Najt & Weiner, 1992 (cited from Hopkin, 1997:30,31; cited from Deharveng, 2004:422,424; cited from Deharveng, Bedos & Weiner, 2011:40) are an extant example of the final completion of the evolutionary trend in the fusion of the occiputo-prothoracic tergites in Poduromorpha. Traditionally, in the Caputanurininae, the headcapsule "is fused with the prothoracic tergite". Stebaeve (1988) and Najt & Weiner (1992) consider this an extraordinary hyperevolution of a Pseudachorutinae lineage (cited from Deharveng, 2004:422), possibly via the very similar Pseudachorutinae (but with not fused protergite): Koreanurina Najt & Weiner, 1992 (cited from Deharveng, Bedos & Weiner, 2011:40), while Lee (1983) gave the genus subfamily status (Caputanurininae). These three genera represent different degrees in head-prothorax fusion, from separate in Koreanurina to completely fused in Leenurina, challenging the validity of the subfamily Caputanurininae as currently defined (cited from Deharveng, Bedos & Weiner, 2011:40).

We challenge this extraordinary hyperevolution of head-prothorax fusion with our hypothesis of the more generalised head-acroprotergite fusion in the Collembola and consider the Caputanurininae as living proof of the ultimate evolutionary step in Poduromorpha by total reduction of the protergite. They share this feature with the Neocollembola. Therefore, the Caputanurininae can be considered as the extant descendants of a lineage of Collembola inbetween Poduromorpha and Neocollembola.

Enlargement of the mesotergum in Entomobryomorpha s.l.

The dorsal area of the prothorax may have fused with the anterior part of the mesotergum, forming a large composite pro-mesotergal sclerite (Christiansen & Bellinger, 1980:15) (Carpentier, 1947, 1949; Bretfeld, 1963; François, 1996 cited from Bitsch & Bitsch, 2000:138-139).

In contrast, based on a comparison of the prothoracic musculature with a basic thoracic musculature model, we presume that the protergite and mesotergite are not or only partly fused, but that at least the mesotergal acrotergite enlarged. In any case, a clear tendency to enlarge the mesotergum can be observed. We will illustrate our presumption with two case studies: Isotomurus maculatus (Isotomidae), based on Hutasse-Jeannenot 1974:93, and Pogonognathellus longicornis (Tomoceridae), based on Carpentier, 1949:48,49.


Fig.3. Isotomurus maculatus
Prothoracic musculature
After Hutasse-Jeannenot, F. 1974.
In Isotomurus maculatus, Hutasse-Jeannenot 1974:93(fig.1) draws four dorsoventral prothoracic muscles that appear to be intersegmental:
from the anterior part of the prosternum to the posterior part of the cephalic tergum (1),
from the anterior part of the prosternum to the median part of the mesotergum (2),
from the posterior part of the prosternum to the posterior part of the cephalic tergum (3),
from the posterior part of the prosternum to the median part of the mesotergum (4).

In our opinion, muscle (3) corresponds with the oblique promeso-intersegmental muscle (green), the ventral end attached to the prosternal apophyse/antecosta and the dorsal end attached to the mesotergal antecosta (see the basic thoracic musculature model of fig.m2). Muscle (4) corresponds with the intrasegmental dorsoventral mesotergosternal muscle (magenta), the ventral end attached to the mesosternal apophyse/antecosta, and the dorsal end attached to the middle of the mesotergite.
It is clear that the mesotergal antecosta is displaced posteriorly and that the mesotergal acrotergite correspondingly is enlarged.


Fig.2. Pogonognathellus longicornis
Thoracic musculature
After Carpentier, F. 1949.
Carpentier (1949) indicates, without any further specification, that, in Tomocerus plumbeus, the prothorax and the mesothorax are fused, dorsally. While ventrally, the three thoracic segments are definitely heteronomous (Carpentier, 1949:48(fig.4),49).
In our opinion, a set of four dorsoventral prothoracic muscles can be recognised that corresponds with those as described in the case of Hutasse-Jeannenot (fig.2). Idem dito the posteriorly displaced mesotergal antecosta (see arrow) and also the enlarged mesotergal acrotergite.
Note the 'knee' in the mesotergal longitudinal muscles (fig.2), which we interprete as the location of the mesotergal antecosta, and therefore the anterior part of the muscles are in fact the protergal longitudinal muscles that enter the head capsule.


To conclude: the enlargement of the mesotergum is mainly due the enlargement of the acrotergite of the mesotergite. This implies that the antecosta of the mesotergite must be displaced more posteriorly (see arrow in fig.2). The tendency to enlarge the meso-acrotergite reaches its climax in Lepidocyrtus in which, in some species, the meso-acrotergite forms a prominent cap that spans the head capsule, such as in Lepidocyrtus paradoxus.

In the context of our labioprotergal fusion hypothesis, two pathways for the enlargement of the acrotergite can be envisaged:

The latter assumption might clarify the fact that in many Entomobryomorpha the mesotergal chaetotaxy is much more complex than in Poduromorpha. In entomobryids and tomocerids, the fusion of the posteroprotergite with the meso-acrotergite gave rise to a more complex chaetotaxy of the mesotergite: with at least one additional anterior row of setae. So the most basic mesotergal setal pattern is then: p' - a - m - p (the anterior p'-row being the p-row of the ancestral protergite). A study of the chaetotaxy in stage 1 of several Isotomidae species also supports the hypothesis of a fusion between pro- en mesonotum (Deharveng in Bitsch & Bitsch, 2000:139). But this is not consistent with the prothoracic musculature found in Entomobryomorpha by Hutasse-Jeannenot 1974:93, and by Carpentier, 1949:48,49. Therefore, we assume that the complex mesotergal chaetotaxy in Entomobryomorpha is the result of polychaetosis and/or plurichaetosis.

Labiotergal fusion in Neelipleona, Entomobryomorpha s.l. and Symphypleona

Based on the two previously described case studies on Isotomurus maculatus (Isotomidae), by Hutasse-Jeannenot 1974:93, and on Pogonognathellus longicornis (Tomoceridae), by Carpentier, 1949:48,49, we conclude that at least in Entomobryomorpha s.l., but presumably also in Neelipleona and Symphypleona, the posteroprotergite eventually fused with the posterior part of the head capsule. This can be deduced from the existance of muscle (2) that corresponds with the posterior intrasegmental lateral dorsoventral protergosternal muscle (blue), the ventral end attached to the anterior part of the prosternite and the dorsal end attached to the end of the cephalic tergum (see the basic thoracic musculature model of fig.m2).

Conclusion

Fig.1. Ordinal phylogenetic relationships
<-pC-+----------------- Poduromorpha
     |
     +--N-+------------ Neelipleona
	  |
	  +------------ Entomobryomorpha s.l. 
	  |
	  +------------ Symphypleona
The ancestral protocollembolans (pC) are characterised by a reduced protergite: the protergite split transversally into an anteroprotergite and a posteroprotergite, of which the anteroprotergite fused with the tergite of the posterior segment of the head capsule.
In the plesiomorph descendants of the protocollembolans, the Poduromorpha, the posteroprotergite is well-developed, generally bearing one transversal row of setae (exeptions: Anurida, Onychiuridae).
In the apomorph descendants of the protocollembolans, called here the Neocollembola (N) (= Neelipleona + Entomobryomorpha s.l. + Symphypleona), the posteroprotergite reduced and fused with the posterior part of the head capsule.
While in Entomobryomorpha s.l. a clear tendency to enlarge the mesotergum can be observed, in Symphypleona, the mesotergum is reduced. The enlargement of the mesotergum in Entomobryomorpha s.l. is mainly due the enlargement of the acromesotergite.

Acknowledgments

I thank Dr Christiansen for reviewing the manuscript and for his constructive comments.

References