73
groups is also supported by the obviously independent
evolution of brooding in these taxa (Osburn 1950 ; Taylor
1988 ; Ostrovsky and Taylor 2005 ; see also Chap. 2 ), suggesting that ctenostomes are paraphyletic and cheilostomes are
polyphyletic (but see Waeschenbach et al. 2012 ).
Similarly, uniserial ctenostome brooders (e.g. Paludicella )
could have inherited a SNP from a uniserial broadcasting
ancestor. Unfortunately, there are no data concerning SNPs
or ITOs in uniserial cheilostomes as yet.
1.3.9.6 Critical Assessment of the Hypothesis
One can argue against this hypothesis, however. There are
both broadcasting species with an ITO and brooders with a
SNP within the same ctenostome taxon Victorella . Thus, on
the one hand, linking the loss of the ITO to the evolution of
brooding is supported. Interestingly, the brooding ctenostome Tanganella muelleri , which has a SNP, shows similar
behaviour to the related brooding victorellid Bulbella
abscondita with a reduced ITO; its polypide bends ventrally
when attaching eggs to its introvert (Braem 1951 ). On the
other hand, all Victorella species form mainly diffuse uniserial chains of zooids. Thus, the example of victorellid ctenostomes does not support the suggestion that the ITO evolved
in a multiserial colony. The same two-chambered ciliated
structure of the ITO in ctenostomes and cheilostomes also
provides evidence for a single origin.
If this is true, the above ideas would need to be reconsidered. In this case, multiserial broadcasters could have
inherited the ITO from uniserial broadcasting ancestors, then
mostly losing it when brooding evolved. The incidence of egg
swallowing may be overestimated too, as most of observations
on spawning were not made under natural fl ow conditions.
On the other hand, uniseriality in cheilostomes could have
evolved secondarily from a multiserial condition many times.
Thus, uniserial broadcasters with an ITO cannot be considered
as fi nal evidence against the hypothesis presented. The ITO
could indeed have been inherited from multiserial broadcasting ancestors.
Evolution of the ITO could be correlated with the prevention of intracoelomic embryo development by delayed
activation of internally fertilized oocytes. The ITO proximal
chamber has a glandular structure, and zygotes are typically
retained within the ITO for variable but brief periods of
time, leading Temkin ( 1994 ) to posit chemical stimulation
of egg activation. If so, then brooders could secondarily lose
this mechanism in favour of some other. In theory, egg activation could be induced by mechanical deformation of the
zygote during release or by contact with sea water in different species (also discussed in Temkin 1996 ). In contrast,
intracoelomic cleavage has been incidentally recorded in
both broadcasting and brooding cheilostomes (see above),
implying that egg activation was not triggered by external
chemical or physical infl uences in such cases.
1.4
Future Research Directions
Despite more than two and a half centuries of observations
and a large number of studied species (see Ostrovsky et al.
2008 , and Appendix I), our knowledge of sexual reproduction in phylum Bryozoa is still inadequate and incomplete,
being based on relatively scarce, disparate information.
Another problem is that, apart from this book and a few relatively recent publications, most available information was
recorded by early naturalists and some of it needs to be
checked. In addition to the general need for new data and the
extension of research to as many taxa as possible, several
important problems still await attention.
1.4.1 Early Gonado- and Gametogenesis
Although several authors described early female cells in
Gymnolaemata, identifying oogonia with certainty is very
challenging, calling into question the reliability of early
observations. Most early descriptions are either superfi cial,
not specifying oogonial stages, or simply confusing.
Theoretically, after differentiation from the mesothelial cell
of the fi rst polypide bud or the cystid wall, the primordial
germ cell(s) should divide, producing two primary oogonia
(each). It is possible that such division was mentioned by
Calvet ( 1900 ) and Chrétien ( 1958 ). In cheilostomes, one of
the oogonial pair will further divide to form the fi rst oocyte
doublet, whereas the second will give rise to two oogonia
(see Sect. 1.2.4 ). However, we do not know about the number
of early female cells and their destiny (degeneration, growth
and divisions). Also the product of the fi rst division of the
primordial germ cell and of the primary oogonium can easily
be confused with the early oocyte doublet in light microscopy. We also have no data concerning how new oogonia (if
any) appear in the mature ovary. TEM studies of both developing and fully formed ovaries should be helpful in answering these questions.
1.4.2 Site of Gonad Origin and Final Location
According to the accounts of different authors, the source
of germ cells is local proliferation and dedifferentiation of
the peritoneum. Early female cells can be moved from their
site of origin towards the fi nal position of the ovary, and
this appears to be facilitated by the developing funicular
network. The site of origin is mainly connected with the
developing polypide bud, although it is also reported on the
cystid wall (Grant 1827 ; Vigelius 1884b ; Pergens 1889 ;
Hageman 1983 ). The fi nal position of the ovary is on the
polypide (the caecum or nearby), the funicular cord(s) or
1.4 Future Research Directions
groups is also supported by the obviously independent
evolution of brooding in these taxa (Osburn 1950 ; Taylor
1988 ; Ostrovsky and Taylor 2005 ; see also Chap. 2 ), suggesting that ctenostomes are paraphyletic and cheilostomes are
polyphyletic (but see Waeschenbach et al. 2012 ).
Similarly, uniserial ctenostome brooders (e.g. Paludicella )
could have inherited a SNP from a uniserial broadcasting
ancestor. Unfortunately, there are no data concerning SNPs
or ITOs in uniserial cheilostomes as yet.
1.3.9.6 Critical Assessment of the Hypothesis
One can argue against this hypothesis, however. There are
both broadcasting species with an ITO and brooders with a
SNP within the same ctenostome taxon Victorella . Thus, on
the one hand, linking the loss of the ITO to the evolution of
brooding is supported. Interestingly, the brooding ctenostome Tanganella muelleri , which has a SNP, shows similar
behaviour to the related brooding victorellid Bulbella
abscondita with a reduced ITO; its polypide bends ventrally
when attaching eggs to its introvert (Braem 1951 ). On the
other hand, all Victorella species form mainly diffuse uniserial chains of zooids. Thus, the example of victorellid ctenostomes does not support the suggestion that the ITO evolved
in a multiserial colony. The same two-chambered ciliated
structure of the ITO in ctenostomes and cheilostomes also
provides evidence for a single origin.
If this is true, the above ideas would need to be reconsidered. In this case, multiserial broadcasters could have
inherited the ITO from uniserial broadcasting ancestors, then
mostly losing it when brooding evolved. The incidence of egg
swallowing may be overestimated too, as most of observations
on spawning were not made under natural fl ow conditions.
On the other hand, uniseriality in cheilostomes could have
evolved secondarily from a multiserial condition many times.
Thus, uniserial broadcasters with an ITO cannot be considered
as fi nal evidence against the hypothesis presented. The ITO
could indeed have been inherited from multiserial broadcasting ancestors.
Evolution of the ITO could be correlated with the prevention of intracoelomic embryo development by delayed
activation of internally fertilized oocytes. The ITO proximal
chamber has a glandular structure, and zygotes are typically
retained within the ITO for variable but brief periods of
time, leading Temkin ( 1994 ) to posit chemical stimulation
of egg activation. If so, then brooders could secondarily lose
this mechanism in favour of some other. In theory, egg activation could be induced by mechanical deformation of the
zygote during release or by contact with sea water in different species (also discussed in Temkin 1996 ). In contrast,
intracoelomic cleavage has been incidentally recorded in
both broadcasting and brooding cheilostomes (see above),
implying that egg activation was not triggered by external
chemical or physical infl uences in such cases.
1.4
Future Research Directions
Despite more than two and a half centuries of observations
and a large number of studied species (see Ostrovsky et al.
2008 , and Appendix I), our knowledge of sexual reproduction in phylum Bryozoa is still inadequate and incomplete,
being based on relatively scarce, disparate information.
Another problem is that, apart from this book and a few relatively recent publications, most available information was
recorded by early naturalists and some of it needs to be
checked. In addition to the general need for new data and the
extension of research to as many taxa as possible, several
important problems still await attention.
1.4.1 Early Gonado- and Gametogenesis
Although several authors described early female cells in
Gymnolaemata, identifying oogonia with certainty is very
challenging, calling into question the reliability of early
observations. Most early descriptions are either superfi cial,
not specifying oogonial stages, or simply confusing.
Theoretically, after differentiation from the mesothelial cell
of the fi rst polypide bud or the cystid wall, the primordial
germ cell(s) should divide, producing two primary oogonia
(each). It is possible that such division was mentioned by
Calvet ( 1900 ) and Chrétien ( 1958 ). In cheilostomes, one of
the oogonial pair will further divide to form the fi rst oocyte
doublet, whereas the second will give rise to two oogonia
(see Sect. 1.2.4 ). However, we do not know about the number
of early female cells and their destiny (degeneration, growth
and divisions). Also the product of the fi rst division of the
primordial germ cell and of the primary oogonium can easily
be confused with the early oocyte doublet in light microscopy. We also have no data concerning how new oogonia (if
any) appear in the mature ovary. TEM studies of both developing and fully formed ovaries should be helpful in answering these questions.
1.4.2 Site of Gonad Origin and Final Location
According to the accounts of different authors, the source
of germ cells is local proliferation and dedifferentiation of
the peritoneum. Early female cells can be moved from their
site of origin towards the fi nal position of the ovary, and
this appears to be facilitated by the developing funicular
network. The site of origin is mainly connected with the
developing polypide bud, although it is also reported on the
cystid wall (Grant 1827 ; Vigelius 1884b ; Pergens 1889 ;
Hageman 1983 ). The fi nal position of the ovary is on the
polypide (the caecum or nearby), the funicular cord(s) or
1.4 Future Research Directions