with eversed ventral tube vesicles
2006 © McClarin, J.
| Checklist of the Collembola: Note on the Ordinal Morphogenetic Relationships of Collembola |
Frans Janssens, Department of Biology, University of Antwerp, Antwerp, B-2020, Belgium
Abstract.
The Collembola evolved into two main groups:
in Neocollembola, with fully developed furca, and
in Poduromorpha, with reduced furca.
The Neocollembola evolved according to 2 schemes:
1. tomoceromorph linear forms, with enlarged third abdominal segment, and
2. panentomobryomorph forms, with enlarged fourth abdominal segment.
The Panentomobryomorpha specialised further into
Euentomobryomorpha, linear comma-shaped forms, and
Symphypleona, subglobular forms with enlarged abdomen.
Neelipleona, subglobular forms with reduced abdomen and enlarged thorax
are a remnant sistergroup of Euentomobryomorpha.
Poduromorpha, such as Neanuridae, Onychiuridae and Tullbergiidae,
with vestigial furcula,
Euentomobryomorpha, such as Microfalculidae and some Isotomidae,
and Symphypleona, such as Mackenziellidae,
with a furcula are secondarely regressive forms.
<-pC-+----+----------- "protoCollembola" (+)
| |
| +----------- Poduromorpha
|
| +----------- Tomoceromorpha
| |
+-Nc-+ +-- Euentomobryomorpha
| +---+
| | +-- Neelipleona
+-Pe-+
+------ Symphypleona
|
This ordinal tree (Fig. 1) is compiled from the views of relationships among orders of Collembola based on phylogenies proposed by Cassagnau (1971), Massoud (1971, 1976) Moen & Ellis (1984), Bretfeld (1986), Fjellberg (1994), Soto-Adames (1996), D'Haese (2002, 2003), Park (2002), Deharveng (2004), Gao & al. (2008), and Xiong & al. (2008). Traditionally, the Collembola have been divided into five groups (Poduromorpha, Metaxypleona, Neelipleona, Entomobryomorpha, and Symphypleona) which different authors have considered to represent orders, sections or every category in between these two. More recently, D'Haese (2002) and Xiong & al. (2008) have concluded that Entomobryomorpha is paraphyletic and D'Haese (2002:1148) proposed Tomoceromorpha as a new basic group of Collembola.
Protocollembola. The hypothetical ancestors of all Collembola, the protoCollembola, are characterised by 1. the protergite being distinctly present, and 2. the furcal precursor muscles of the third abdominal segment being present. The protoCollembola and the direct plesiomorph descendants are extinct.
pT+ pT+ pT+
f3+ f3+ f3+
<--pC--o--pC'--o--pC"--o protoCollembola (+)
| |
| | pT+
| | f3-
| +-------o Poduromorpha
|
| pT-
| f3+
+-------o Neocollembola (Nc)
|
Poduromorpha.
In the Poduromorpha, the apomorph descendants of the protoCollembola',
the furcal precursor muscles of the third abdominal segment are absent (f3-).
In the plesiomorph descendants of the protoCollembola', the protoCollembola",
the furcal precursor muscles of the third abdominal segment remain
present (f3+). These plesiomorph descendants are all extinct (+).
The Poduromorpha are characterised by:
1. the protergite being distinctly present, and
2. an elongated and segmented habitus.
These characters are assumed to be plesiomorphies
(after Thibaud, Schulz & da Gama Assalino, 2004:3).
The combination of characters
present in members of the Isotogastruridae seem to represent a transitional
stage between the orders Poduromorpha and Entomobryomorpha.
The ordinal position of Isotogastruridae
is resolved by D'Haese (2003:577,578) as sister group of Tullbergiidae.
To avoid breaking the taxonomic hierarchy,
it is here tentatively placed as a 'familia incertae sedis'
in Isotogastruroidea, close to Onychiuroidea.
Metaxypleona. Contrary to Moen & Ellis (1984) and Stebaeva (1988, cited from D'Haese, 2003:564), but more in line with Soto-Adames (1996), Carpenter & Wheeler (1999), D'Haese (2002, 2003), Park (2002) and Deharveng (2004), Metaxypleona is here not maintained to give Podura an ordinal taxonomic status. We do not agree with Börner (1906), who pointed out that Podura has some synapomorphies with Symphypleona: in particular the hypognath head and the long curved dentes of the furca. We believe that these characters developed independently in both groups and are simply the result of homoplasy. Handschin (1929:19) suggested to place Podura inbetween the Arthropleona (Poduromorpha and Entomobryomorpha s.l.) and Symphypleona. Salmon (1964:100) defined Metaxypleona1 for Podura and Actaletes as a sister group of Symphypleona. This author saw Podura as a very early stage in the morphological processes of specialisation which lead ultimately to the Symphypleona. Contrary to Salmon's opinion, we will show that Podura and Symphypleona are paraphyletic groups. Moen & Ellis (1984:202-203) redefined Metaxypleona, by retaining Poduridae as the only member of the order. In addition, the authors described four putative synapomorphies that support a sister group relationship between Metaxypleona and Symphypleona-Neelipleona. However, taken into account the morphological study of Frish (1978) on the heart in Podura, Actaletes and representatives of Poduromorpha, Symphypleona, and Entomobryomorpha, the Metaxypleona is here considered in close relationship with the Poduromorpha. Frish concluded that the heart of poduromorphs, including Podura, has six ostia, while the higher entomobryomorphs have five ostia, and in Symphypleona the ostia are reduced to three. The epicuticular ultrastructural lattice of microtubercles of Podura appears to be more like that of Poduromorpha than of Symphypleona (Dallai & Malatesta, 1973:135-139). According to Hemmer (1990), Podura aquatica has a haploid chromosome number of n=11, the highest of any Collembola (Hopkin, 1997:34). Symphypleona typically have n=5 up to n=7 (Dallai & al., 2000:56). This makes a close relationship between Metaxypleona and Symphypleona questionable (Hopkin, 1997:34). Jordana & Arbea (1997:15) point out that the retrocerebral complex of Podura is identical to that of Hypogastruridae, indicating that Metaxypleona and Poduromorpha are more closely related. Deharveng (2000:in litt.) notes that Podura is very near to Hypogastrura by its chaetotaxy. We agree with D'Haese (2002, 2003), Park (2002) and Xiong & al. (2008) that Podura is well embedded into Poduromorpha. Therefore, Metaxypleona is rejected, and in line with Palacios-Vargas (1994:409), Poduroidea is tentatively reestablished to give Podura superfamiliar status within the Poduromorpha, close to Hypogastruroidea.
Tomoceromorpha.
Xiong & al. (2008) confirm the findings of D'Haese (2002) that the
superfamily Tomoceroidea is monophyletic and that Entomobryomorpha is
paraphyletic.
Unfortunately, their statement
"Some special morphological characters of
Tomoceroidea have been identified: the third abdominal segment
longer than the fourth segment, the body with scales, and the
antennae with four segments. These characters are very different
from those of the other groups in Entomobryomorpha, but similar
to those of some species in Poduromorpha (Hopkin, 1997)."
is a misconception. The characters "the body with scales, and the antennae
with four segments" are not at all different from those of other groups in
Entomobryomorpha. Scales also are present in Entomobryidae and in Paronellidae.
On the contrary, scales are absent in Poduromorpha. And in all Collembola
the antennae have 4 segments.
The monophyly of Tomoceroidea is weakly supported (63.6% in average)
(Xiong & al. 2008).
The monophyly of Poduromorpha+Tomoceroidea is not supported
(<50% in the Bayesian tree) (Xiong & al. 2008).
Therefore, we place the new ordo Tomoceromorpha D'Haese 2002:1148
tentativelly in the Neocollembola.
Tomoceromorpha are characterised by:
1. an absent protergite, and
2. a muscular springing mechanism mainly situated in the third
abdominal segment, making the third abdominal segment larger than the fourth.
Entomobryomorpha. D'Haese (2002:1148) concludes that Entomobryomorpha is paraphyletic. Recently, Xiong & al. (2008) have confirmed that Entomobryomorpha is paraphyletic and Tomoceroidea has been proposed as a new basic group of Collembola. Therefore Entomobryomorpha is here rejected.
Panentomobryomorpha. The sister group of Tomoceromorpha, tentatively called the Panentomobryomorpha (Fig. 1:Pe), are characterised by: 1. an absent protergite, and 2. a muscular springing mechanism mainly situated in the fourth abdominal segment, making the fourth abdominal segment larger than the third.
|
Symphypleona.
The Symphypleona are characterised by:
1. an absent protergite,
2. a muscular springing mechanism mainly situated in the fourth
abdominal segment, making the fourth abdominal segment larger than the third,
3. a ventral tube having two long eversible vesicles (see Fig. 2), and
4. a subglobular habitus.
Contrary to the morphological phylogeny of D'Haese (2003:571), that supports
the traditionally accepted
Neelipleona-Symphypleona sister group relationschip,
on the basis of two
synapomorphies: 1. antennae inserted on the side of the frontal area, without
contact with the clypeal area, and 2. a body of globular shape, with fused
thoracic and abdominal segments,
we support the paraphyly of Neelipleona-Symphypleona as shown by the combined
morphological and molecular phylogeny of Robertson (2001).
Note that the molecular phylogeny of D'Haese (2002:1148) defines Symphypleona
as being paraphyletic.
Xiong & al. (2008) supported the monophyly of Symphypleona
moderately (72.3% in average).
The globular body shape is a very weak synapomorphy since it is
realised in two completely different and even opposite ways.
In Neelidae, the thorax makes up the greater part of the trunk, while in
Sminthuridae s.l. the abdomen is the largest body region
(Christiansen & Bellinger, 1998:1175).
In Neelipleona, the globular shape of the body is realised by strongly
reduced abdominal segments I-IV, which are fused and strongly curved
dorsoventrad,
in combination with a relatively enlarged metathoracic segment.
In Symphypleona, the globular body shape is realised by
enlarged abdominal segments I-IV,
blown-up to such an extent that the stretched intersegmental conjuctivae
are more or less indistinct.
Xiong & al. (2008) tested the grouping of Neelides with
the Symphypleona in a ribosomal RNA phylogeny, and found respectively
weak Maximum Likelihood (ML) support (68/0.99),
weak Bayesian support (62/0.91),
and no Maximum Parsimony (MP) support (<50).
Taking into account the MP, ML and Bayesian trees, in average, support of
the grouping of Neelides with Symphypleona is < 60, in other words the
monophyly of Neelipleona+Symphypleona is not supported.
Therefore, we consider the conclusion of Xiong & al. (2008:734)
"Neelipleona as the sister group to Symphypleona" not supported.
In line with Christiansen & Bellinger (1998:1175) we support the
independant convergent derivation of Neelidae and Sminthuridae s.l. from
their respective arthropleone ancestors,
i.e. Symphypleona from poduromorph ancestors,
and Neelipleona from entomobryomorph ancestors.
Euentomobryomorpha.
The Euentomobryomorpha are characterised by:
1. an absent protergite,
2. a muscular springing mechanism mainly situated in the fourth
abdominal segment, making the fourth abdominal segment larger than the third,
3. a ventral tube having two short eversible vesicles, which is considered to
be a plesiomorphy, since it is also found in Poduromorpha and Tomoceromorpha,
and
4. a smooth midgut.
The Euentomobryomorpha comprise the remaining entomobryomorph superfamilies
Isotomoidae and Entomobryoidea.
Neelipleona.
The Neelipleona are characterised by:
1. an absent protergite,
2. a muscular springing mechanism mainly situated in the fourth abdominal
segment, making the fourth abdominal segment larger than the third
(the fourth abdominal segment is developed strongest (Willem, 1900:65)),
3. a ventral tube having two short eversible vesicles, and
4. a diverticulate midgut.
Taking into account the discussion of the globular body shape
and in addition that
1. the form of the retinaculum
of Neelidae corresponds with that of arthropleone Collembola versus
that of Sminthuridae (Börner, 1906:4);
2. abdominal bothriotricha, typical for Sminthuridae s.l., are lacking
in Neelidae (Börner, 1906:4),
making a close relationship between Neelidae and Sminthuridae s.l.
questionable;
3. the anterior setal row of the clypeus, the prelabral setae of Yosii 1976,
only has 2+2 setae as in arthropleone Collembola (Bretfeld, 1999:16);
4. the postembryonal development of Neelipleona is more in line with that of the
arthropleone collembolans than with the Symphypleona, such as the longer
lifetime, the number of adult instars and the alternating productive and
non-reproductive instars (Blancquaert & Mertens, 1979:129);
5. the male genital organ of Megalothorax is much simpler than
that of Sminthurus due to the abscence of the prolonged deferens
channels and the specialised successive regions in the gonads; their structure
resembles that of the arthropleone Collembola, with an additional longitudinal
condensation (Willem, 1900:67); and
6. the peculiar sex-determination mechanism, as well as the consequent aberrant
spermatogenesis, is common to all members of the Symphypleona;
this process has been demonstrated in seven genera
from three different families: Sminthuridae and Dicyrtomidae
(Dallai et al. 1999, 2000 cited from Dallai et al. 2001:238)
and Bourletiellidae (Dallai et al. 2001:237-238);
on the contrary, it has not been found in members of the Neelipleona, nor in
other Collembolan species (Dallai et al. 1999 cited from Dallai et al. 2001:238).
Therefore, tentatively, we put Neelipleona as sistergroup of the
Euentomobryomorpha.
Gao et al. (2008:1142) confirmed the close relationship between
Entomobryomorpha and Neelipleona in a phylogenetic rDNA study of basal Hexapoda.
 |
 |
The hypothetical protocollembolan ancestral body segmentation is modeled
according to the segmental proportions as can be observed
in juvenile extant Collembola.
Additionally, it is presumed that in the Protocollembola
1. the prothorax is partially reduced;
2. the furca is not yet fully developed; initially, it is small: a furcula.
3. the metaabdomen is small compared to the pro- and mesoabdomen.
The smaller size of the metaabdomen is related to the
smaller sized hindgut.
The upwards rotated position of the metaabdomen is related to
the ventro-posterior insertion of the furcula.
Biomechanical model of the mesoabdominal furcal tension
The development of the furca is closely related
with the development of the furcal muscles and the mesoabdomen.
The antero-posterior tension caused by the furcal muscles
on the mesoabdominal segments eventually
result in a dorso-ventrally abdominal curvature.
This tension is primarely caused by the flexor muscles of the furca.
The furcal muscles of the hypothetical protocollembolan
are modelled as three sets of muscles:
1. the flexor furcae inflex the furca underneath the trunk in its
resting position,
2. the extensor furcae and
3. the levator furcae extend the furca behind the trunk to effectuate
the jump.
The flexor and extensor muscles are derived from muscles that operated
the ancestral abdominal legs: respectively
the tergal promotor muscles and the tergal remotor muscles.
The levator muscles are derived from the
tergosternal muscles and intersegmental muscles of the abdominal segments.
The levator muscles operate the furca indirectly by creating a
hydrostatic pressure in the abdominal and furcal lumen.
There are two sets of flexor furcae, one at each side of the mesoabdomen.
Each lateral set can comprise up to two sets of muscles:
a primary segmental set and a secondary intersegmental set.
Both are attached ventrally to the antero-lateral rim of the furcal base.
The segmental set is attached dorsally
to the dorso-lateral antecosta of the posterior mesoabdominal tergite.
The intersegmental set is attached dorsally
to the dorso-lateral antecosta of the anterior mesoabdominal tergite,
such as in Orchesella cincta cf. Schaller (1970:32,Fig.50).
The intersegmental set develops secondarely and only in case of large furca
(such as found in Neocollembola).
There are two sets of extensor furcae, one at each side of the mesoabdomen.
The dorso-ventral extensor furcae are attached
dorsally to the dorso-lateral antecosta of the posterior mesoabdominal tergite
and ventrally to the postero-lateral rim of the furcal base.
There are two sets of dorso-ventral levator furcae.
Both sets are attached medio-ventrally at the antecosta
of the posterior mesoabdominal sternite.
One segmental set is attached medio-dorsally to the center of the
posterior mesoabdominal tergite.
The intersegmental set is attached medio-dorsally to the antecosta
of the anterior metaabdominal tergite.
Only the intersegmental flexor muscles contribute to the
mesoabdominal antero-posterior tension.
At the moment of docking the furca in the retinaculum,
this tension is maximum.
The flexor tension creates a downwards momentum
on the posterior mesoabdominal segment.
Evolutionary, to compensate for this tension,
the mesoabdominal terga tend to enlarge
while the mesoabdominal sterna tend to reduce.
The elongation of the mesoabdominal terga has been realised differently
in Tomoceromorpha and Panentomobryomorpha.
In Tomoceromorpha, the anterior mesoabdominal tergum, the third abdominal
tergum, is enlarged, while in Panentomobryomorpha, the posterior
mesoabdominal tergum, the fourth abdominal tergum, is enlarged.
For a discussion on the evolution of the furcal muscles,
see Janssens (2008).
The furca itself is modelled in analogy with the sequence of primitive limb articulation of the hexapod ground plan according to Kukalova-Peck (1992, 2008): 1. epicoxa, 2. subcoxa, 3. coxa, 4. trochanter, 5. prefemur, 6. femur, 7. patella, 8. tibia, 9. basitarsus, 10. eutarsus, 11. posttarsus (1992) or pretarsus (2008). Kukalova-Peck claims that "These segments are all unequivocally identified by their own muscle insertions; they occur serially in all limb-derived appendages on the head, thorax, and abdomen."
| Ancestral hexapod limb articulation | Collembolan limb articulation | Collembolan furca articulation | |
|---|---|---|---|
| 1. | epicoxa | epicoxa | manubrium |
| 2. | subcoxa | subcoxa | |
| 3. | coxa | coxa | |
| 4. | trochanter | trochanter | 5. | prefemur | femur | dens |
| 6. | femur | ||
| 7. | patella | tibia, with pseudotarsus and unguis | mucro |
| 8. | tibia | ||
| 9. | basitarsus | empodial tubercle with optional unguiculus | 10. | eutarsus |
| 11. | posttarsus or pretarsus |
Based on the homology described in Table I, the protocollembolan manubrium is modelled as a syncoxa with 4 pair-wise fused articles. The tergal condyles of the fused epicoxae are modelled as the basal condyles of the manubrium. It is assumed that the sternal epicoxal condyles became redundant due to the epicoxal fusion and therefore are reduced completely. The redundant condyles of the fused subcoxae are reduced completely. The outer condyles of the fused coxae remain functional while the inner ones reduce. The outer condyles of the trochanter reduce due to the fusion with the coxa. The result is a protocollembollan manubrium with a dicondylic basis and apex. Due to the fusion, the sternal remotors and promotors become redundant and are reduced. The dorsal remotors become furcal extensors and the dorsal promotors become furcal flexors. Three sets of extensor and flexor muscles remain functional: the epicoxal, the subcoxal, and the coxal set. The trochanteral set reduces completely. The epicoxal and subcoxal set cooperate to rotate the basal manubrium backward and forward. The coxal set rotates the distal dentes backward and forward. The syncoxal manubrium as well as the dentes can rotate in their respective condyles up to an angle of about 90 degrees. This allows the furca to be extended up to about 180 degrees. Due to the double set of condyles, the epicoxal and coxal condyles, the furca can be extended 180 degrees in the time required to extend the manubrial base 90 degrees. Due to the double set of manubrial condyles, the furca can be extended at such a high speed that is otherwise not possible.
The compact protocollembolan body shape reduces the possibility of
ever finding a fossil collembolan of similar proportions to gigantic
early millipedes which reached nearly 2m (Hopkin, 1997:26).
The furca is the most important factor that constrains
the absolute maximum body size.
The largest extant Collembola are found among the
Tetrodontophorinae (up to 9 mm),
onychiurid poduromorphans with a strongly shortened furca
and Uchidanurinae (up to 17 mm),
neanurid poduromorphans with completely vestigial furca.
 |
 |
The enlargement of the furca developed synchroneous with the enlargement of the furcal muscles and the development of the secondary intersegmental flexor muscles. The larger furcal muscles have a direct impact on the overall shape of the body: to accomodate the larger muscles, the relative size of the abdomen increased. To compensate for the longitudinal strain, caused by the furcal flexor muscles, the mesoabdominal terga tend to elongate while the mesoabdominal sterna tend to shorten.
 |
 |
In the plesiomorph descendents of the Panentomobryomorpha, to compensate for the increased antero-posterior tension of the intersegmental furcal flexor muscles, the trunk started to curve dorsoventrad.
In one group, the Euentomobryomorpha,
due to the strong proabdomen the dorsoventrad curvature is
relatively small, resulting in a comma-shaped lateral habitus.
Note that as a secondarely derived character, in the Euentomobryomorpha
(more in particular, in the Microfalculidae and in some Isotomidae),
the furca can be vestigial up to absent,
reducing correspondingly the trunk curvature.
 |
 |
The rather straightforward ordinal phylogeny that is proposed here is based on a dichotomic phylogeny of the progressively increasing function of the furca. It is assumed that in the protocollembola the furca is not well developed. Initially, it is small: a furcula. The Poduromorpha is the oldest group that specialised this small furca. The furcula reduced even further in size in stead of that it developed. In some species it even disappeared. However, others specialised in increasing the furcula: the 'higher' Collembola, the Neocollembola, characterised with a vestigial protergite. The larger furca has an effect on the overall shape of the body: it becomes dorso-ventrally curved due to the large antero-posterior muscles that spring-load the furca. The larger furca is realised according to two different bodyplans. In the first scheme, the larger furcal flexing muscles are accomodated by the third abdominal segment. In these Tomoceromorpha, the body curvature is minimal. In the second scheme, the Panentomobryomorpha, the larger furcal flexing muscles are accomodated by the fourth abdominal segment. In the Euentomobryomorpha, the curvature is typically resulting in a comma-shaped lateral habitus. Note that as a derived character, in the Euentomobryomorpha, the furca can be reduced up to absent, reducing correspondingly the body curvature. The Neelipleona with a strongly curved body are here considered as a remnant sistergroup of the Euentomobryomorpha. In the Symphypleona the body curvature reaches its maximum development: the last thoracic segments and first four abdominal segments are curved upwards to such an extent that they resemble the shape of an inverse 'U', reducing effectively the body length.
Notes
1 From the Greek 'metaxy' meaning 'found in the middle'
or 'between'.