248
size of eggs and embryos is the main obstacle for using dry
mass in studies of matrotrophy in most invertebrates; the only
study that used such data dealt with a terrestrial isopod
(Lawlor 1976 ). The same obstacle pertains to bryozoans. In
this phylum, apart from ultrastructural evidence (see above),
embryo enlargement during incubation, together with accompanying morphological changes in the embryophore and the
embryo, are currently the main criteria used to ascertain the
occurrence of EEN. Water uptake is obviously a useful criterion but the degree to which it takes place is unknown. It may
be noted that some increase in embryo volume was recorded
in a number of non-matrotrophic cheilostome brooders. In the
vast majority it ranged from 1.05 to 1.3-fold, reached 2.5-fold
in a few species (see Table 1.8 ). Such an increase is comparable with that recorded in species with presumed incipient
matrotrophy, but there were neither a developed embryophore
nor detectable changes in embryo cells to provide evidence of
nutrient transfer in the former species, suggesting water
uptake in them (see also Ostrovsky 2013 ).
3.3.2 Multiple Origins of Placentotrophy
in Cheilostomata
Matrotrophy and, in particular, placentotrophy are generally
regarded as having evolved many times in different classes
of vertebrates (Wourms 1981 ; Blackburn et al. 1985 ;
Blackburn 1992 , 1999b , 2005a ; Wooding and Burton 2008 ).
Similarly, the distribution of the patterns of sexual reproduction across Bryozoa strongly suggests that placentotrophy
evolved independently in all three bryozoan classes and
within both gymnolaemate orders (Ostrovsky et al. 2009a ).
Unfortunately, a robust phylogenetic framework is still lacking for Bryozoa. Published molecular phylogenies are very
incomplete (at best analysing species from less than 10% of
all described genera) and contradictory in many important
details (see Tsyganov-Bodounov et al. 2009 ; Fuchs et al.
2009 ; Knight et al. 2011 ; Waeschenbach et al. 2012 ). As for
matrotrophic cheilostomes, only a few species from the
matrotrophic genera Bugula , Beania , Watersipora and
Cellaria as well as Bicellariella ciliata and Celleporella hyalina were involved in the molecular analysis, and in all published phylogenetic trees the distribution of matrotrophic
taxa is very patchy. Species of Bugula and Beania are situated within the same branch in the trees made by Knight
et al. ( 2011 ), and Bicellariella ciliata is placed in the same
branch with Bugula in the tree by Tsyganov-Bodounov et al.
( 2009 ). Such a placement implies the possibility of a common ancestor with EEN for some lineages within Buguloidea
but a more rigorous analysis of the superfamily is required,
involving many more taxa.
One of the major arguments in favour of this suggestion
for Cheilostomata is the presence of two, or sometimes three,
patterns of sexual reproduction in the same families and the
presence of two patterns in the same genera. In other words,
in many instances closely related species can be matrotrophic
or non-matrotrophic, or, if matrotrophic, may have different
modes of oogenesis. Species with pattern II (non-matrotrophic with macrolecithal oogenesis) and pattern IV (matrotrophic with macrolecithal oogenesis) have been recorded in the
families Candidae, Cribrilinidae and Hippothoidae. Patterns
II, III (matrotrophic with microlecithal oogenesis) and IV are
known in Bugulidae, Flustridae, Cellariidae and, apparently,
Catenicellidae. Two different patterns have been found in
Gregarinidra (II and III), Isosecurifl ustra (II and IV) and
Microporidae and Cellaria (III and IV) (Ostrovsky et al.
2009a ; Ostrovsky 2013 ). A similar situation has been
described in teleost fi shes of the families Poeciliidae and
Zenarchopteridae, in which close relatives “vary either in a
presence or absence of matrotrophy or in the degree to which
matrotrophy is developed” (Reznick et al. 2002 , 2007a , p.
2570). To note, the molecular analysis showed that matrotrophy may have evolved independently not only within these
families but also within several different genera (Reznick
et al. 2002 , 2007a ; Pollux et al. 2009 ; Pires et al. 2011 ;
Meredith et al. 2011 ).
The cheilostome genus Bugula is notable for having various degrees of matrotrophy resulting in embryo enlargement
from 6.3- to 500-fold in different species (Woollacott and
Zimmer 1975 ; Dyrynda and King 1983 ; pers. obs.). This
attribute is reminiscent of the continuum of variation in
matrotrophic provisioning recorded in such fi sh genera as
Poeciliopsis , Nomorhamphus and Dermogenys (see Reznick
et al. 2002 , 2007a ). Variation in the degree of EEN has been
suggested among populations of the poeciliid fi sh Phalloceros
caudimaculatus (see Arias and Reznick 2000 ).
Intraspecifi c variation among oocyte types and larval
increase during matrotrophic incubation were detected in
distant populations of Bugula fl abellata . Dyrynda and King
( 1983 , p. 489) described “telolecithal” (= macrolecithal) eggs
77 μm in diameter in the Irish Sea colonies that had larval size
150 μm. My material from New Zealand contained oligolecithal eggs (96 × 55 μm) and larvae 160 × 120 μm in diameter.
Even if these populations are represented by different (but
clearly very closely related) species, the presence of two
different modes of oogenesis may point to the shift between
patterns IV and III within this clade (Ostrovsky 2013 ).
3.3.3 Plausibility of an Alternative Scenario
The proposed sequence of events in the evolution of placentotrophy within cheilostome bryozoans may be questioned
by making a case for reversibility of matrotrophy. An alternative scenario would then be that EEN originated in a hypothetical early brooder producing relatively small mesolecithal
3 Evolution of Reproductive Patterns in Cheilostomata
size of eggs and embryos is the main obstacle for using dry
mass in studies of matrotrophy in most invertebrates; the only
study that used such data dealt with a terrestrial isopod
(Lawlor 1976 ). The same obstacle pertains to bryozoans. In
this phylum, apart from ultrastructural evidence (see above),
embryo enlargement during incubation, together with accompanying morphological changes in the embryophore and the
embryo, are currently the main criteria used to ascertain the
occurrence of EEN. Water uptake is obviously a useful criterion but the degree to which it takes place is unknown. It may
be noted that some increase in embryo volume was recorded
in a number of non-matrotrophic cheilostome brooders. In the
vast majority it ranged from 1.05 to 1.3-fold, reached 2.5-fold
in a few species (see Table 1.8 ). Such an increase is comparable with that recorded in species with presumed incipient
matrotrophy, but there were neither a developed embryophore
nor detectable changes in embryo cells to provide evidence of
nutrient transfer in the former species, suggesting water
uptake in them (see also Ostrovsky 2013 ).
3.3.2 Multiple Origins of Placentotrophy
in Cheilostomata
Matrotrophy and, in particular, placentotrophy are generally
regarded as having evolved many times in different classes
of vertebrates (Wourms 1981 ; Blackburn et al. 1985 ;
Blackburn 1992 , 1999b , 2005a ; Wooding and Burton 2008 ).
Similarly, the distribution of the patterns of sexual reproduction across Bryozoa strongly suggests that placentotrophy
evolved independently in all three bryozoan classes and
within both gymnolaemate orders (Ostrovsky et al. 2009a ).
Unfortunately, a robust phylogenetic framework is still lacking for Bryozoa. Published molecular phylogenies are very
incomplete (at best analysing species from less than 10% of
all described genera) and contradictory in many important
details (see Tsyganov-Bodounov et al. 2009 ; Fuchs et al.
2009 ; Knight et al. 2011 ; Waeschenbach et al. 2012 ). As for
matrotrophic cheilostomes, only a few species from the
matrotrophic genera Bugula , Beania , Watersipora and
Cellaria as well as Bicellariella ciliata and Celleporella hyalina were involved in the molecular analysis, and in all published phylogenetic trees the distribution of matrotrophic
taxa is very patchy. Species of Bugula and Beania are situated within the same branch in the trees made by Knight
et al. ( 2011 ), and Bicellariella ciliata is placed in the same
branch with Bugula in the tree by Tsyganov-Bodounov et al.
( 2009 ). Such a placement implies the possibility of a common ancestor with EEN for some lineages within Buguloidea
but a more rigorous analysis of the superfamily is required,
involving many more taxa.
One of the major arguments in favour of this suggestion
for Cheilostomata is the presence of two, or sometimes three,
patterns of sexual reproduction in the same families and the
presence of two patterns in the same genera. In other words,
in many instances closely related species can be matrotrophic
or non-matrotrophic, or, if matrotrophic, may have different
modes of oogenesis. Species with pattern II (non-matrotrophic with macrolecithal oogenesis) and pattern IV (matrotrophic with macrolecithal oogenesis) have been recorded in the
families Candidae, Cribrilinidae and Hippothoidae. Patterns
II, III (matrotrophic with microlecithal oogenesis) and IV are
known in Bugulidae, Flustridae, Cellariidae and, apparently,
Catenicellidae. Two different patterns have been found in
Gregarinidra (II and III), Isosecurifl ustra (II and IV) and
Microporidae and Cellaria (III and IV) (Ostrovsky et al.
2009a ; Ostrovsky 2013 ). A similar situation has been
described in teleost fi shes of the families Poeciliidae and
Zenarchopteridae, in which close relatives “vary either in a
presence or absence of matrotrophy or in the degree to which
matrotrophy is developed” (Reznick et al. 2002 , 2007a , p.
2570). To note, the molecular analysis showed that matrotrophy may have evolved independently not only within these
families but also within several different genera (Reznick
et al. 2002 , 2007a ; Pollux et al. 2009 ; Pires et al. 2011 ;
Meredith et al. 2011 ).
The cheilostome genus Bugula is notable for having various degrees of matrotrophy resulting in embryo enlargement
from 6.3- to 500-fold in different species (Woollacott and
Zimmer 1975 ; Dyrynda and King 1983 ; pers. obs.). This
attribute is reminiscent of the continuum of variation in
matrotrophic provisioning recorded in such fi sh genera as
Poeciliopsis , Nomorhamphus and Dermogenys (see Reznick
et al. 2002 , 2007a ). Variation in the degree of EEN has been
suggested among populations of the poeciliid fi sh Phalloceros
caudimaculatus (see Arias and Reznick 2000 ).
Intraspecifi c variation among oocyte types and larval
increase during matrotrophic incubation were detected in
distant populations of Bugula fl abellata . Dyrynda and King
( 1983 , p. 489) described “telolecithal” (= macrolecithal) eggs
77 μm in diameter in the Irish Sea colonies that had larval size
150 μm. My material from New Zealand contained oligolecithal eggs (96 × 55 μm) and larvae 160 × 120 μm in diameter.
Even if these populations are represented by different (but
clearly very closely related) species, the presence of two
different modes of oogenesis may point to the shift between
patterns IV and III within this clade (Ostrovsky 2013 ).
3.3.3 Plausibility of an Alternative Scenario
The proposed sequence of events in the evolution of placentotrophy within cheilostome bryozoans may be questioned
by making a case for reversibility of matrotrophy. An alternative scenario would then be that EEN originated in a hypothetical early brooder producing relatively small mesolecithal
3 Evolution of Reproductive Patterns in Cheilostomata