254
many times as incubation emerged. Thus, a shift in oogenesis
towards the origination of a non-feeding larval type was
always associated with (presumably preceded) the evolution
of parental care. If, instead, lecithotrophy has a monophyletic origin among Cheilostomata, then the above arguments
should be reconsidered, and one should expect the existence
of malacostegan-like cheilostomes without brooding but
with non-feeding larvae. Such variants are not yet known,
however, and the case of Arbocuspis bellula [formerly
Electra ], forming large eggs but considered to be a malacostegan (Marcus 1938 ), requires further study.
Larval feeding is not known in either Phylactolaemata or
Stenolaemata, thus obscuring the question of whether a nonfeeding larva evolved independently in these classes or was
inherited from their ancestors. As for the class Gymnolaemata,
it seems that lecithotrophy evolved numerous times in both
its orders, Ctenostomata (Sect. 3.4.4 ) and Cheilostomata. If
the independent evolution of brooding is a marker for the
evolution of endotrophy, then in the cheilostome suborders
Inovicellina (having external membranous brood sacs) and
Scrupariina (having both external membranous sacs in skeletal bivalve ovicells) lecithotrophic larvae presumably
evolved independently in both clades. Judging from their
morphology, these suborders may have had different
malacostegan- like (i.e. non-brooding with planktotrophic
larvae) ancestors, although Inovicellina might also have
originated from a ctenostome ancestor [polyphyly of
Cheilostomata is demonstrated in the study by Jebram
( 1992 )]. Signifi cantly, the genera Scruparia (Scrupariina)
and Aetea (Inovicellina) group with malacostegans in a
molecular study by Waeschenbach et al. ( 2012 ). Also, the
non-feeding larva of Scruparia chelata is strongly reminiscent of the shelled cyphonautes-like larva of the ctenostome
Flustrellidra hispida but lacks the shell (see Barrois 1877 ;
Zimmer and Woollacott 1977a , b ).
Families Eucrateidae and Leiosalpingidae [both members
of suborder Scrupariina in Gordon ( 2012 )] brood their
embryos in external membranous sacs, similar to the situation in Aetea , while the supposedly related Scrupariidae have
bivalved ovicells. Since external sacs evolved several times
in both ctenostomes and cheilostomes, there is no obvious
connection between Aetea , Eucratea and leiosalpingids.
Similarly, the structure of the ovicell in the Scrupariidae differs from the conventional ovicells of other cheilostomes and
most probably evolved independently. Overall, it appears
that embryonic incubation evolved independently (twice?) in
Scrupariina.
Other examples include the cheilostome families
Calloporidae, Tendridae, Belluloporidae, Thalamoporellidae
and Alysidiidae: the structure of their cystids is easily derived
from that in malacostegans (directly or via intermediates),
and their brood chambers give evidence that these are nonhomologous. Thus, if these groups independently evolved
from the different malacostegan ancestors, the only group
known to have reproductive pattern I, then lecithotrophy
originated in them independently too. Incidentally, the
revealed topologies in two variants of the molecular analysis
made by Knight et al. ( 2011 ) indirectly confi rm the independent origins of thalamoporellids (and their relatives, steginoporellids) and calloporids from malacostegans. Also,
according to Marcus ( 1939 ), the non-feeding larva of
Thalamoporella evelinae is only reminiscent of the larva of
Scruparia chelata – another cheilostome that seems to have
evolved brooding and lecithotrophy independently (see also
Zimmer and Woollacott 1977a ).
Important arguments supporting the hypothesis of multiple and independent origins of lecithotrophy in cheilostome
bryozoans are (1) large time gaps between the apparent origins of groups with endotrophic larvae (as evidenced by the
fossil record), i.e. Calloporidae, Albian (Middle Cretaceous);
Scrupariidae,
Maastrichtian
(Late
Cretaceous);
Thalamoporellidae, Eocene; Belluloporidae, Pleistocene;
Tendridae, Recent; and (2) the absence of direct phylogenetic connections between these taxa. I suggest that all of
these groups evolved from different malacostegine ancestors
(with cyphonautes larvae) and that lecithotrophy and brooding originated in them independently. If the endotrophic
larva evolved only once, it follows that all of these groups
evolved from a hypothetical malcostegan-like clade with
lecithotrophy but without embryonic incubation. Moreover,
a further inference may be drawn that this clade had to survive from the Late Cretaceous until the present day. If this
scenario were correct, one would expect at least some such
cheilostomes (with a lecithotrophic larva but without brooding) to have survived. A possible candidate is the previously
mentioned Arbocuspis bellula , an electrid malacostegan that
should be a broadcaster but is said to produce a large egg
(Marcus 1938 ) and may in fact be an internal brooder (see
Chap. 1 ). In their molecular analysis (which included data
from GenBank), Knight et al. ( 2011 ) found that this species
associated with species of Electra .
But even if lecithotrophy evolved once in the major cheilostome lineage Flustrina (=Neocheilostomina), currently
understood as monophyletic, anatomical data show that
brooding evolved numerous times within this lineage.
Accordingly, the new suborders Tendrina, Thalamoporellina
and Belluloporina, three new superfamilies, Tendroidea,
Thalamoporelloidea and Belluloporoidea, and the corresponding family Belluloporidae are established herein (see
Appendix II for diagnoses). The case of Alysidiidae requires
additional study.
3.4.1.2 Tendra zostericola
Returning to the question of when embryonic incubation
evolved in respect to the origin of lecithotrophy, it should
additionally be noted that there are two opposing hypotheses
3 Evolution of Reproductive Patterns in Cheilostomata
many times as incubation emerged. Thus, a shift in oogenesis
towards the origination of a non-feeding larval type was
always associated with (presumably preceded) the evolution
of parental care. If, instead, lecithotrophy has a monophyletic origin among Cheilostomata, then the above arguments
should be reconsidered, and one should expect the existence
of malacostegan-like cheilostomes without brooding but
with non-feeding larvae. Such variants are not yet known,
however, and the case of Arbocuspis bellula [formerly
Electra ], forming large eggs but considered to be a malacostegan (Marcus 1938 ), requires further study.
Larval feeding is not known in either Phylactolaemata or
Stenolaemata, thus obscuring the question of whether a nonfeeding larva evolved independently in these classes or was
inherited from their ancestors. As for the class Gymnolaemata,
it seems that lecithotrophy evolved numerous times in both
its orders, Ctenostomata (Sect. 3.4.4 ) and Cheilostomata. If
the independent evolution of brooding is a marker for the
evolution of endotrophy, then in the cheilostome suborders
Inovicellina (having external membranous brood sacs) and
Scrupariina (having both external membranous sacs in skeletal bivalve ovicells) lecithotrophic larvae presumably
evolved independently in both clades. Judging from their
morphology, these suborders may have had different
malacostegan- like (i.e. non-brooding with planktotrophic
larvae) ancestors, although Inovicellina might also have
originated from a ctenostome ancestor [polyphyly of
Cheilostomata is demonstrated in the study by Jebram
( 1992 )]. Signifi cantly, the genera Scruparia (Scrupariina)
and Aetea (Inovicellina) group with malacostegans in a
molecular study by Waeschenbach et al. ( 2012 ). Also, the
non-feeding larva of Scruparia chelata is strongly reminiscent of the shelled cyphonautes-like larva of the ctenostome
Flustrellidra hispida but lacks the shell (see Barrois 1877 ;
Zimmer and Woollacott 1977a , b ).
Families Eucrateidae and Leiosalpingidae [both members
of suborder Scrupariina in Gordon ( 2012 )] brood their
embryos in external membranous sacs, similar to the situation in Aetea , while the supposedly related Scrupariidae have
bivalved ovicells. Since external sacs evolved several times
in both ctenostomes and cheilostomes, there is no obvious
connection between Aetea , Eucratea and leiosalpingids.
Similarly, the structure of the ovicell in the Scrupariidae differs from the conventional ovicells of other cheilostomes and
most probably evolved independently. Overall, it appears
that embryonic incubation evolved independently (twice?) in
Scrupariina.
Other examples include the cheilostome families
Calloporidae, Tendridae, Belluloporidae, Thalamoporellidae
and Alysidiidae: the structure of their cystids is easily derived
from that in malacostegans (directly or via intermediates),
and their brood chambers give evidence that these are nonhomologous. Thus, if these groups independently evolved
from the different malacostegan ancestors, the only group
known to have reproductive pattern I, then lecithotrophy
originated in them independently too. Incidentally, the
revealed topologies in two variants of the molecular analysis
made by Knight et al. ( 2011 ) indirectly confi rm the independent origins of thalamoporellids (and their relatives, steginoporellids) and calloporids from malacostegans. Also,
according to Marcus ( 1939 ), the non-feeding larva of
Thalamoporella evelinae is only reminiscent of the larva of
Scruparia chelata – another cheilostome that seems to have
evolved brooding and lecithotrophy independently (see also
Zimmer and Woollacott 1977a ).
Important arguments supporting the hypothesis of multiple and independent origins of lecithotrophy in cheilostome
bryozoans are (1) large time gaps between the apparent origins of groups with endotrophic larvae (as evidenced by the
fossil record), i.e. Calloporidae, Albian (Middle Cretaceous);
Scrupariidae,
Maastrichtian
(Late
Cretaceous);
Thalamoporellidae, Eocene; Belluloporidae, Pleistocene;
Tendridae, Recent; and (2) the absence of direct phylogenetic connections between these taxa. I suggest that all of
these groups evolved from different malacostegine ancestors
(with cyphonautes larvae) and that lecithotrophy and brooding originated in them independently. If the endotrophic
larva evolved only once, it follows that all of these groups
evolved from a hypothetical malcostegan-like clade with
lecithotrophy but without embryonic incubation. Moreover,
a further inference may be drawn that this clade had to survive from the Late Cretaceous until the present day. If this
scenario were correct, one would expect at least some such
cheilostomes (with a lecithotrophic larva but without brooding) to have survived. A possible candidate is the previously
mentioned Arbocuspis bellula , an electrid malacostegan that
should be a broadcaster but is said to produce a large egg
(Marcus 1938 ) and may in fact be an internal brooder (see
Chap. 1 ). In their molecular analysis (which included data
from GenBank), Knight et al. ( 2011 ) found that this species
associated with species of Electra .
But even if lecithotrophy evolved once in the major cheilostome lineage Flustrina (=Neocheilostomina), currently
understood as monophyletic, anatomical data show that
brooding evolved numerous times within this lineage.
Accordingly, the new suborders Tendrina, Thalamoporellina
and Belluloporina, three new superfamilies, Tendroidea,
Thalamoporelloidea and Belluloporoidea, and the corresponding family Belluloporidae are established herein (see
Appendix II for diagnoses). The case of Alysidiidae requires
additional study.
3.4.1.2 Tendra zostericola
Returning to the question of when embryonic incubation
evolved in respect to the origin of lecithotrophy, it should
additionally be noted that there are two opposing hypotheses
3 Evolution of Reproductive Patterns in Cheilostomata