235
and metamorphosis, the structures ensuring its autonomous
feeding (capture and digestion of food particles) would have
been no longer needed. The free-swimming period was considerably shortened for the same reason as recorded for larvae of most of the known incubating bryozoans that swim
freely for less than 24 h, in comparison with planktotrophic
larvae that live for periods of 1 week to 2 months (Dudley
1973 ; Yoshioka 1982 ; Cook 1985 ). Also, a reduction of this
period and the evolution of embryonic incubation (which is
compulsory for the development of endotrophic larvae in
bryozoans) might explain the loss of larval protective structures, that is, the shell of the cyphonautes. Similar changes
have been described in sea urchins, sea stars and brittle stars
(Wray and Raff 1991 ; Byrne 1991a ; Wray 1992 ; McEdward
and Janies 1993 ; Raff 1996 ). A detailed analysis of the loss
of the food- capturing structures in connection with the
acquisition of large oocytes and the transition to a non-feeding larva in some sedentary polychaetes can be found in the
work of Pernet ( 2003 ).
Though our knowledge of bryozoan larvae is incomplete
and fragmentary (reviewed in Barrois 1877 ; Ryland 1974 ,
1976 ; Zimmer and Woollacott 1977a ; Cook 1985 ; Reed
1987 , 1991 ; Mukai et al. 1997 ), we know enough to be able
to say that the class Gymnolaemata, with its broad range of
larval forms, illustrates the above-described hypothetical
sequence of the transition to lecithotrophy. Facultative planktotrophic larvae have not been described in Bryozoa, but in
this phylum there are species with lecithotrophic larvae completely lacking a gut and species whose larvae have a nonfunctioning digestive tract reduced to varying degrees.
Non-feeding larvae in three such species in the order
Ctenostomata, i.e. Flustrellidra hispida , Pherusella tubulosa
and P . brevituba , have retained the gut, which is incomplete
posteriorly, and a bivalve shell homologous to that of the
cyphonautes (Zimmer and Woollacott 1977a ). The larva of
the ctenostome Triticella fl ava looks very much like a
cyphonautes but lacks the shell, a mouth and, apparently, an
anus. However, it is the only non-feeding larva to develop a
vestibulum, a body wall invagination characteristic of
cyphonautes larvae. After a short external brooding phase,
the development of such larvae is completed in the plankton,
lasting altogether for a week. According to Ström ( 1969 ),
fully formed larvae further survived in an aquarium for a
month, decreasing in size during this time, indicating that
these larvae use their internal resources (see also Zimmer
and Woollacott 1977a ). In addition, according to Repiachoff
( 1875 , 1878 ) and Ostroumoff ( 1886b ), the coronate larva of
the cheilostome Tendra zostericola has a non-functioning
rudimentary gut, consisting of midgut and rectum (data on
oesophagus and mouth require checking). On the other hand,
the vast majority of bryozoan larvae lack any trace of feeding
and protective structures.
These examples show that the transition from a
planktotrophic to a lecithotrophic larval type is sometimes
accompanied by partial loss of the structures that enable
feeding and protection of the larva, thus illustrating a gradual
transition from one type to another. A rapid transition cannot
be excluded, however. In any case, such a reduction appears
to be expedient only if the larva no longer needs to feed on
its own and has a shorter free-swimming period. According
to the assessment made by Strathmann ( 1978a ) on the basis
of data in the literature, planktotrophy was lost in Bryozoa
three to six times (see also below).
The above facts and arguments are completely at odds with
the hypothesis, suggested by Silén ( 1944 ), that the cyphonautes
larva is of secondary origin. It was based on the assumption
that brood chambers in the phylum are homologous. Silén
thought that the oldest among them was “embryonary”, viz the
internal brood sac of Phylactolaemata, and that the structures
responsible for embryonic incubation in the “CheiloCtenostomata” evolved from it. Silén thus argued that Recent
bryozoans lacking brood chambers lost the capacity to brood,
which, in turn, resulted in modifi cation of the larva. Silén
postulated that this transition occurred within the CheiloСtenostomata several times, and that broadcasters evolved
rather late. Recently, Fuchs et al. ( 2011 , p. 11) presented
data on gene-expression patterns indicating “that planktonic
larvae might have secondarily evolved in bryozoans”.
As mentioned in Sect. 2.4.2 , Santagata and Banta ( 1996 ,
p. 178) proposed a hypothesis according to which “vestibular
brooding preceded evolution of ovicells among cheilostomes”. An outcome of vestibular incubation was loss of the
planktotrophic larva. In contrast with my hypothesis, these
authors suggested that extraembryonic nutrition via the
hypertrophied vestibular epithelium was responsible for
enlargement of the embryo and the shift to endotrophy.
Changes in oogenesis were not mentioned. Overall, their
hypothesis was based on misinterpreted facts and assumptions and cannot be considered probable (see also Ostrovsky
2002 ; Taylor and McKinney 2002 ; Ostrovsky et al. 2006 ).
Although planktotrophy in invertebrates does seem to
have evolved secondarily in a number of cases (see McHugh
and Rouse 1998 ; Collin 2004 ; Collin et al. 2007 ), data on the
evolution of brooding and reproductive patterns in cheilostome bryozoans, as well as the sequence in which the major
clades appeared in the fossil record, render Silén’s hypothesis as purely speculative and based on assumptions not facts.
Reproductive pattern I is indeed the rarest among Bryozoa.
However, suborder Malacostegina in which it occurs is
the oldest cheilostome clade, and the morphology of the
cyphonautes larva corresponds to the structure of the trochophore, considered to be the initial larval morphotype in many
groups of marine invertebrates (Cori 1941 ; Jägersten 1972 ;
Strathmann 1978a ; Ivanova-Kazas 1986 ). The presence of a
3.1 Modifi cation of Oogenesis and Its Evolutionary Consequences
and metamorphosis, the structures ensuring its autonomous
feeding (capture and digestion of food particles) would have
been no longer needed. The free-swimming period was considerably shortened for the same reason as recorded for larvae of most of the known incubating bryozoans that swim
freely for less than 24 h, in comparison with planktotrophic
larvae that live for periods of 1 week to 2 months (Dudley
1973 ; Yoshioka 1982 ; Cook 1985 ). Also, a reduction of this
period and the evolution of embryonic incubation (which is
compulsory for the development of endotrophic larvae in
bryozoans) might explain the loss of larval protective structures, that is, the shell of the cyphonautes. Similar changes
have been described in sea urchins, sea stars and brittle stars
(Wray and Raff 1991 ; Byrne 1991a ; Wray 1992 ; McEdward
and Janies 1993 ; Raff 1996 ). A detailed analysis of the loss
of the food- capturing structures in connection with the
acquisition of large oocytes and the transition to a non-feeding larva in some sedentary polychaetes can be found in the
work of Pernet ( 2003 ).
Though our knowledge of bryozoan larvae is incomplete
and fragmentary (reviewed in Barrois 1877 ; Ryland 1974 ,
1976 ; Zimmer and Woollacott 1977a ; Cook 1985 ; Reed
1987 , 1991 ; Mukai et al. 1997 ), we know enough to be able
to say that the class Gymnolaemata, with its broad range of
larval forms, illustrates the above-described hypothetical
sequence of the transition to lecithotrophy. Facultative planktotrophic larvae have not been described in Bryozoa, but in
this phylum there are species with lecithotrophic larvae completely lacking a gut and species whose larvae have a nonfunctioning digestive tract reduced to varying degrees.
Non-feeding larvae in three such species in the order
Ctenostomata, i.e. Flustrellidra hispida , Pherusella tubulosa
and P . brevituba , have retained the gut, which is incomplete
posteriorly, and a bivalve shell homologous to that of the
cyphonautes (Zimmer and Woollacott 1977a ). The larva of
the ctenostome Triticella fl ava looks very much like a
cyphonautes but lacks the shell, a mouth and, apparently, an
anus. However, it is the only non-feeding larva to develop a
vestibulum, a body wall invagination characteristic of
cyphonautes larvae. After a short external brooding phase,
the development of such larvae is completed in the plankton,
lasting altogether for a week. According to Ström ( 1969 ),
fully formed larvae further survived in an aquarium for a
month, decreasing in size during this time, indicating that
these larvae use their internal resources (see also Zimmer
and Woollacott 1977a ). In addition, according to Repiachoff
( 1875 , 1878 ) and Ostroumoff ( 1886b ), the coronate larva of
the cheilostome Tendra zostericola has a non-functioning
rudimentary gut, consisting of midgut and rectum (data on
oesophagus and mouth require checking). On the other hand,
the vast majority of bryozoan larvae lack any trace of feeding
and protective structures.
These examples show that the transition from a
planktotrophic to a lecithotrophic larval type is sometimes
accompanied by partial loss of the structures that enable
feeding and protection of the larva, thus illustrating a gradual
transition from one type to another. A rapid transition cannot
be excluded, however. In any case, such a reduction appears
to be expedient only if the larva no longer needs to feed on
its own and has a shorter free-swimming period. According
to the assessment made by Strathmann ( 1978a ) on the basis
of data in the literature, planktotrophy was lost in Bryozoa
three to six times (see also below).
The above facts and arguments are completely at odds with
the hypothesis, suggested by Silén ( 1944 ), that the cyphonautes
larva is of secondary origin. It was based on the assumption
that brood chambers in the phylum are homologous. Silén
thought that the oldest among them was “embryonary”, viz the
internal brood sac of Phylactolaemata, and that the structures
responsible for embryonic incubation in the “CheiloCtenostomata” evolved from it. Silén thus argued that Recent
bryozoans lacking brood chambers lost the capacity to brood,
which, in turn, resulted in modifi cation of the larva. Silén
postulated that this transition occurred within the CheiloСtenostomata several times, and that broadcasters evolved
rather late. Recently, Fuchs et al. ( 2011 , p. 11) presented
data on gene-expression patterns indicating “that planktonic
larvae might have secondarily evolved in bryozoans”.
As mentioned in Sect. 2.4.2 , Santagata and Banta ( 1996 ,
p. 178) proposed a hypothesis according to which “vestibular
brooding preceded evolution of ovicells among cheilostomes”. An outcome of vestibular incubation was loss of the
planktotrophic larva. In contrast with my hypothesis, these
authors suggested that extraembryonic nutrition via the
hypertrophied vestibular epithelium was responsible for
enlargement of the embryo and the shift to endotrophy.
Changes in oogenesis were not mentioned. Overall, their
hypothesis was based on misinterpreted facts and assumptions and cannot be considered probable (see also Ostrovsky
2002 ; Taylor and McKinney 2002 ; Ostrovsky et al. 2006 ).
Although planktotrophy in invertebrates does seem to
have evolved secondarily in a number of cases (see McHugh
and Rouse 1998 ; Collin 2004 ; Collin et al. 2007 ), data on the
evolution of brooding and reproductive patterns in cheilostome bryozoans, as well as the sequence in which the major
clades appeared in the fossil record, render Silén’s hypothesis as purely speculative and based on assumptions not facts.
Reproductive pattern I is indeed the rarest among Bryozoa.
However, suborder Malacostegina in which it occurs is
the oldest cheilostome clade, and the morphology of the
cyphonautes larva corresponds to the structure of the trochophore, considered to be the initial larval morphotype in many
groups of marine invertebrates (Cori 1941 ; Jägersten 1972 ;
Strathmann 1978a ; Ivanova-Kazas 1986 ). The presence of a
3.1 Modifi cation of Oogenesis and Its Evolutionary Consequences