245
development of the nucleus is one of the main arguments in
this connection. For example, in Porella minuta and P . smitti ,
mature nurse cells have a very large nucleus occupying most
of the cell, with the cytoplasm looking like a narrow peripheral ring. Large nuclei indicate that these cells actively produce RNA, although their cytoplasm also contains yolk
granules. This example may represent an intermediate evolutionary stage from the ancient variant (the nurse cell predominantly producing yolk) to the advanced variant (the nurse
cell forming ribosomes). Other species possibly illustrating
this trend are Hippoporina reticulatopunctata and Bugulopsis
monotrypa , in which mature nurse cells do not contain yolk
granules whereas the nurse cells of early vitellogenic oocytes
do, as if the early stages of nurse-cell functioning recapitulated the ancient form of synthesis and the later stages the
advanced one.
3.3
Evolution of Matrotrophic Incubation
in Cheilostomata
3.3.1 Origin of Placentotrophy
In contrast to most other invertebrate phyla, extraembryonic
nutrition (EEN) is common in Bryozoa (Levin and Bridges
1995 ; Batygina et al. 2006 ; Ostrovsky et al. 2009a ; Lidgard
et al. 2012 ). All matrotrophic bryozoans are equipped with
temporary structure(s) that, together with the apposed part of
the embryo, act as a “simplifi ed placenta-like system”
(Woollacott and Zimmer 1972a , b ; 1975 ). EEN is thought to
be obligatory in living stenolaemates and phylactolaemates
(Reed 1991 ; Mukai et al. 1997 ; Ostrovsky 2009 ), relatively
widespread in Cheilostomata (Ostrovsky et al. 2008a , 2009a )
and, as recently shown at the ultrastructural level, present in
Ctenostomata (Ostrovsky and Schwaha 2011 ).
In discussing bryozoan reproductive strategies, Nielsen
( 1990 ) emphasized that as well as the three major patterns
there are also several “intermediate types”, alluding to the
total diversity of bryozoan reproductive variants. In the
event, this terminology is applicable – the recently discovered pattern IV is an intermediate variant between reproductive patterns II and III (Ostrovsky et al. 2009a ). Following
the terminology of Kasyanov ( 1989 ), there was a transition
from a lecithotrophic embryonic strategy to a placental one.
Insofar as the numbering terminology of reproductive patterns I to III has become established in the literature the
newly discovered pattern had to be assigned IV, but this is
not intended to refl ect the evolutionary sequence.
Although cheilostomes with reproductive patterns I and III
share the feature of yolk-poor oocytes, their oogenesis differs
considerably, indicating that pattern III is unlikely to have
evolved from pattern I. For instance, it would be hard to
explain the great difference in the number of oocytes
formed by species with these patterns during oogenesis.
Paleontological data also do not support the idea that species
with pattern III evolved from an ancestor with pattern I. In
contrast, the type, size and number of oocytes in bryozoans
with patterns II and IV are similar, indicating the essential
similarity, if not identity, of their oogenesis types. In addition,
species with these patterns may have more than one vitellogenic doublet in the ovary; further, these patterns are found
within the same genera and families. All these facts support
the idea that pattern IV evolved on the basis of pattern II via
acquisition of the placental analogue, further transforming to
pattern III (Ostrovsky et al. 2009a ; Ostrovsky 2013 ).
A recently proposed scenario describing the main steps of
the advent of placentotrophy in cheilostome bryozoans suggested that the evolution of the new reproductive patterns
proceeded as a cascade of events including transitions from
reproductive pattern I to pattern II, from pattern II to pattern
IV, and further from pattern IV to pattern III (Ostrovsky et al.
2009a ; Ostrovsky 2013 ). These transitions involved two corresponding shifts in oogenesis from oligo- to macrolecithal
(during transition from pattern I to II) and back (from pattern
IV to III). The latter shift could have been triggered by the
acquisition of placentotrophy during incubation, which gradually substituted ovarian vitellogenesis as a major source of
the nutrients needed for embryonic development. An inverse
correlation between the degree of maternal provisioning
during oogenesis and matrotrophic gestation is well-known
among invertebrates and vertebrates. For instance, less-yolky
eggs are known to develop in echinoderms possessing EEN
(Byrne 1991b ; Wray 1995a ; Byrne and Cerra 1996 ; Byrne
et al. 1999 ). Greatly reduced vitelline systems are characteristic of some matrotrophic monogenean fl atworms (Cable
and Tinsley 1991 ). Such reduction is considered to be an
evolutionary trend in matrotrophic cestodes (Swiderski and
Xylander 2000 ; Korneva 2005 and references therein) and
the same trend can be also inferred from the data on egg
types in scorpions (Francke 1982 ) and matrotrophic isopods
(Hoese and Janssen 1989 ). Among vertebrates, some highly
placentotrophic squamate reptiles ovulate eggs with a
reduced egg content (reviewed in Blackburn 1993 ). Finally,
in mammals, the evolution of placentation resulted in a shift
to microlecithal oogenesis based on the loss of the yolk
genes (Rothchild 2003 ; Brawand et al. 2008 ).
Why nutrient transfer during incubation should have
evolved in bryozoans is unclear. One possibility is that initially it was relatively unimportant and played no role in
embryonic development. The next step could have appeared
in the form of precocious ovulation and oviposition, as a
result of a non-mature egg being transported to the incubation chamber. In this way, the role of EEN in provisioning
resources to the embryo may have gradually changed from
supplementary to central. This change is likely to have
accompanied a transition from a weakly functioning (or small)
3.3 Evolution of Matrotrophic Incubation in Cheilostomata
development of the nucleus is one of the main arguments in
this connection. For example, in Porella minuta and P . smitti ,
mature nurse cells have a very large nucleus occupying most
of the cell, with the cytoplasm looking like a narrow peripheral ring. Large nuclei indicate that these cells actively produce RNA, although their cytoplasm also contains yolk
granules. This example may represent an intermediate evolutionary stage from the ancient variant (the nurse cell predominantly producing yolk) to the advanced variant (the nurse
cell forming ribosomes). Other species possibly illustrating
this trend are Hippoporina reticulatopunctata and Bugulopsis
monotrypa , in which mature nurse cells do not contain yolk
granules whereas the nurse cells of early vitellogenic oocytes
do, as if the early stages of nurse-cell functioning recapitulated the ancient form of synthesis and the later stages the
advanced one.
3.3
Evolution of Matrotrophic Incubation
in Cheilostomata
3.3.1 Origin of Placentotrophy
In contrast to most other invertebrate phyla, extraembryonic
nutrition (EEN) is common in Bryozoa (Levin and Bridges
1995 ; Batygina et al. 2006 ; Ostrovsky et al. 2009a ; Lidgard
et al. 2012 ). All matrotrophic bryozoans are equipped with
temporary structure(s) that, together with the apposed part of
the embryo, act as a “simplifi ed placenta-like system”
(Woollacott and Zimmer 1972a , b ; 1975 ). EEN is thought to
be obligatory in living stenolaemates and phylactolaemates
(Reed 1991 ; Mukai et al. 1997 ; Ostrovsky 2009 ), relatively
widespread in Cheilostomata (Ostrovsky et al. 2008a , 2009a )
and, as recently shown at the ultrastructural level, present in
Ctenostomata (Ostrovsky and Schwaha 2011 ).
In discussing bryozoan reproductive strategies, Nielsen
( 1990 ) emphasized that as well as the three major patterns
there are also several “intermediate types”, alluding to the
total diversity of bryozoan reproductive variants. In the
event, this terminology is applicable – the recently discovered pattern IV is an intermediate variant between reproductive patterns II and III (Ostrovsky et al. 2009a ). Following
the terminology of Kasyanov ( 1989 ), there was a transition
from a lecithotrophic embryonic strategy to a placental one.
Insofar as the numbering terminology of reproductive patterns I to III has become established in the literature the
newly discovered pattern had to be assigned IV, but this is
not intended to refl ect the evolutionary sequence.
Although cheilostomes with reproductive patterns I and III
share the feature of yolk-poor oocytes, their oogenesis differs
considerably, indicating that pattern III is unlikely to have
evolved from pattern I. For instance, it would be hard to
explain the great difference in the number of oocytes
formed by species with these patterns during oogenesis.
Paleontological data also do not support the idea that species
with pattern III evolved from an ancestor with pattern I. In
contrast, the type, size and number of oocytes in bryozoans
with patterns II and IV are similar, indicating the essential
similarity, if not identity, of their oogenesis types. In addition,
species with these patterns may have more than one vitellogenic doublet in the ovary; further, these patterns are found
within the same genera and families. All these facts support
the idea that pattern IV evolved on the basis of pattern II via
acquisition of the placental analogue, further transforming to
pattern III (Ostrovsky et al. 2009a ; Ostrovsky 2013 ).
A recently proposed scenario describing the main steps of
the advent of placentotrophy in cheilostome bryozoans suggested that the evolution of the new reproductive patterns
proceeded as a cascade of events including transitions from
reproductive pattern I to pattern II, from pattern II to pattern
IV, and further from pattern IV to pattern III (Ostrovsky et al.
2009a ; Ostrovsky 2013 ). These transitions involved two corresponding shifts in oogenesis from oligo- to macrolecithal
(during transition from pattern I to II) and back (from pattern
IV to III). The latter shift could have been triggered by the
acquisition of placentotrophy during incubation, which gradually substituted ovarian vitellogenesis as a major source of
the nutrients needed for embryonic development. An inverse
correlation between the degree of maternal provisioning
during oogenesis and matrotrophic gestation is well-known
among invertebrates and vertebrates. For instance, less-yolky
eggs are known to develop in echinoderms possessing EEN
(Byrne 1991b ; Wray 1995a ; Byrne and Cerra 1996 ; Byrne
et al. 1999 ). Greatly reduced vitelline systems are characteristic of some matrotrophic monogenean fl atworms (Cable
and Tinsley 1991 ). Such reduction is considered to be an
evolutionary trend in matrotrophic cestodes (Swiderski and
Xylander 2000 ; Korneva 2005 and references therein) and
the same trend can be also inferred from the data on egg
types in scorpions (Francke 1982 ) and matrotrophic isopods
(Hoese and Janssen 1989 ). Among vertebrates, some highly
placentotrophic squamate reptiles ovulate eggs with a
reduced egg content (reviewed in Blackburn 1993 ). Finally,
in mammals, the evolution of placentation resulted in a shift
to microlecithal oogenesis based on the loss of the yolk
genes (Rothchild 2003 ; Brawand et al. 2008 ).
Why nutrient transfer during incubation should have
evolved in bryozoans is unclear. One possibility is that initially it was relatively unimportant and played no role in
embryonic development. The next step could have appeared
in the form of precocious ovulation and oviposition, as a
result of a non-mature egg being transported to the incubation chamber. In this way, the role of EEN in provisioning
resources to the embryo may have gradually changed from
supplementary to central. This change is likely to have
accompanied a transition from a weakly functioning (or small)
3.3 Evolution of Matrotrophic Incubation in Cheilostomata