247
some teleost fi shes (Turner 1940 ; Wourms 1981 ; Reznick
et al. 2002 , 2007a ; Marsh-Matthews et al. 2010 ).
The fi nal step in this hypothetical transition from pattern
IV to pattern III was a shift from the production of small
macrolecithal to meso- or oligolecithal eggs, supported by
substantial placentotrophy. Among species with pattern III,
maximum embryonic enlargement was recorded in those
with the smallest oocytes ( Bugula neritina , Reciprocus regalis , Mollia multijuncta , Pterocella scutella ), and it is in these
species that the difference between egg size and brood-cavity
size was most prominent.
As in pattern IV, species with pattern III demonstrate different degrees of embryo enlargement and embryophore
development. There is no clear correlation between these two
characters, however, and species with both strong and modest
hypertrophy of the cells of the placental analogue demonstrate a wide variation in embryo enlargement. For instance,
species with modest hypertrophy of embryophore cells
showed a range of enlargement from 4.9-fold in C . tenuirostris to 53.4 in Mollia multijuncta , as did species with strong
hypertrophy of these cells: from 6.3-fold in Bugula fl abellata
to 310-fold in B . neritina . Thus, as in pattern IV, two variants
of substantial placentation – with modest and strong hypertrophy of placental-analogue cells – are detectable among
species producing eggs with a small amount of yolk.
The specifi c case of Myriapora truncata , which combines
large macrolecithal eggs and strong hypertrophy of the
embryophore cells (pattern IV), is puzzling. Its zygote occupies the entire cavity of the ovicell so that further embryonic
growth should be strongly restricted. This case may be an
example of rapid evolution of a well-developed placental
analogue, contrasting with the model of gradual acquisition
of the embryophore as discussed above. Another possible
explanation is that in Myriapora truncata the embryophore
serves exclusively for excretory purposes, removing wastes
produced by the large embryo (Ostrovsky 2013 ).
3.3.1.1 Critical Remarks
Conclusions about incipient matrotrophy in bryozoans can
be criticized because of the lack of data on intraspecifi c
(intracolonial, seasonal, geographic) variation in embryophore development and larval size in the species in which
inferred EEN is responsible for embryo “increase”. Working
on living material, Cancino and Hughes ( 1988 ), Wendt
( 2000 ), Marshall et al. ( 2003 ), Marshall and Keough ( 2003 ,
2004a , 2006 , 2008a , b) and Kosman and Pernet ( 2009 )
showed that larval size can vary or be rather stable within and
between populations of the same species. All of the species
studied are matrotrophic brooders belonging to the genera
Celleporella , Bugula and Watersipora , some with small, others
with substantial increase in embryo size during incubation.
It is not clear why larval size varies in these taxa, however.
In Bugula neritina , larvae increase with parent- colony wet
mass (Marshall et al. 2003 ), at higher colony densities
(Marshall and Keough 2008b ) and in colonies following
toxicant exposure (Marshall 2008 ), and diminish in experimentally halved colonies (Marshall and Keough 2004b ),
thus being supposedly dependant on colony state (see also
Marshall and Keough 2009 ). If egg size is stable then
recorded variations in larval size would refl ect variation in
the EEN, i.e. the placenta is a means by which the colony
controls larval size. Such functional fl exibility of bryozoan
placental analogues may point to an evolutionary past in
which their progressive modifi cation led to the acquisition of
substantial matrotrophy. In fact, Bugula neritina demonstrates the largest larval increase during brooding ever
recorded in cheilostomes.
It is important to note that no research was carried out on
the egg size variation in the afore-mentioned studies; thus it
is not known if (and how) this trait might infl uence variation
in larval size. Evidence for such a connection would considerably add to our understanding of variability in larval size.
In my fi xed material, the maximum size of both mature
oocytes and late embryos was stable although the sample
size was low compared with the studies mentioned above,
and volume estimation using histological sections is clearly
not as precise as the methods used by the above authors. It is
clear that larger sample sizes are required to increase the statistical power of embryo-enlargement analyses. However,
multiplication of embryophore cells, their hypertrophy and
cytological change as well as discrepancy in size between
the mature egg and brood cavity and changes in embryonic
yolk content all point to the existence of EEN. Even if the
increase in embryo size is small and nutrient transfer is negligible, the morphological evidence strongly supports the
inference that some exchange (more than just of gases and
water) occurs between the embryo and the parent.
Another factor to consider is that of water absorption as a
reason for embryo enlargement. In studies on vertebrates,
measurements of changes in dry mass are currently the main
indicator of EEN whereas volume and wet mass are not considered as reliable criteria (Blackburn 1994 ). One reason is
that developing embryos always increase in wet mass and
volume (due to water uptake), regardless of whether matrotrophy is present.
In contrast with vertebrates, embryonic size increase is
still widely used as evidence of matrotrophy in invertebrates
and lower chordates. Experiments with radiolabelling and
diet manipulation (Toolson 1985 ; Hoese and Janssen 1989 ;
Frick 1998 ) as well as ultrastructural studies (Domenici and
Gremigni 1977 ; Cable and Tinsley 1991 ; Schwartz and
Dimock 2001 ; Korneva 2005 ) are very rare. Instead many
authors have recorded and described anatomical changes in
both the parent and the offspring during incubation, considering
such changes as additional evidence for matrotrophy (e.g.
Hagan 1951 ; Mukai et al. 1987 ; Farley 2001 ; etc.). The small
3.3 Evolution of Matrotrophic Incubation in Cheilostomata
some teleost fi shes (Turner 1940 ; Wourms 1981 ; Reznick
et al. 2002 , 2007a ; Marsh-Matthews et al. 2010 ).
The fi nal step in this hypothetical transition from pattern
IV to pattern III was a shift from the production of small
macrolecithal to meso- or oligolecithal eggs, supported by
substantial placentotrophy. Among species with pattern III,
maximum embryonic enlargement was recorded in those
with the smallest oocytes ( Bugula neritina , Reciprocus regalis , Mollia multijuncta , Pterocella scutella ), and it is in these
species that the difference between egg size and brood-cavity
size was most prominent.
As in pattern IV, species with pattern III demonstrate different degrees of embryo enlargement and embryophore
development. There is no clear correlation between these two
characters, however, and species with both strong and modest
hypertrophy of the cells of the placental analogue demonstrate a wide variation in embryo enlargement. For instance,
species with modest hypertrophy of embryophore cells
showed a range of enlargement from 4.9-fold in C . tenuirostris to 53.4 in Mollia multijuncta , as did species with strong
hypertrophy of these cells: from 6.3-fold in Bugula fl abellata
to 310-fold in B . neritina . Thus, as in pattern IV, two variants
of substantial placentation – with modest and strong hypertrophy of placental-analogue cells – are detectable among
species producing eggs with a small amount of yolk.
The specifi c case of Myriapora truncata , which combines
large macrolecithal eggs and strong hypertrophy of the
embryophore cells (pattern IV), is puzzling. Its zygote occupies the entire cavity of the ovicell so that further embryonic
growth should be strongly restricted. This case may be an
example of rapid evolution of a well-developed placental
analogue, contrasting with the model of gradual acquisition
of the embryophore as discussed above. Another possible
explanation is that in Myriapora truncata the embryophore
serves exclusively for excretory purposes, removing wastes
produced by the large embryo (Ostrovsky 2013 ).
3.3.1.1 Critical Remarks
Conclusions about incipient matrotrophy in bryozoans can
be criticized because of the lack of data on intraspecifi c
(intracolonial, seasonal, geographic) variation in embryophore development and larval size in the species in which
inferred EEN is responsible for embryo “increase”. Working
on living material, Cancino and Hughes ( 1988 ), Wendt
( 2000 ), Marshall et al. ( 2003 ), Marshall and Keough ( 2003 ,
2004a , 2006 , 2008a , b) and Kosman and Pernet ( 2009 )
showed that larval size can vary or be rather stable within and
between populations of the same species. All of the species
studied are matrotrophic brooders belonging to the genera
Celleporella , Bugula and Watersipora , some with small, others
with substantial increase in embryo size during incubation.
It is not clear why larval size varies in these taxa, however.
In Bugula neritina , larvae increase with parent- colony wet
mass (Marshall et al. 2003 ), at higher colony densities
(Marshall and Keough 2008b ) and in colonies following
toxicant exposure (Marshall 2008 ), and diminish in experimentally halved colonies (Marshall and Keough 2004b ),
thus being supposedly dependant on colony state (see also
Marshall and Keough 2009 ). If egg size is stable then
recorded variations in larval size would refl ect variation in
the EEN, i.e. the placenta is a means by which the colony
controls larval size. Such functional fl exibility of bryozoan
placental analogues may point to an evolutionary past in
which their progressive modifi cation led to the acquisition of
substantial matrotrophy. In fact, Bugula neritina demonstrates the largest larval increase during brooding ever
recorded in cheilostomes.
It is important to note that no research was carried out on
the egg size variation in the afore-mentioned studies; thus it
is not known if (and how) this trait might infl uence variation
in larval size. Evidence for such a connection would considerably add to our understanding of variability in larval size.
In my fi xed material, the maximum size of both mature
oocytes and late embryos was stable although the sample
size was low compared with the studies mentioned above,
and volume estimation using histological sections is clearly
not as precise as the methods used by the above authors. It is
clear that larger sample sizes are required to increase the statistical power of embryo-enlargement analyses. However,
multiplication of embryophore cells, their hypertrophy and
cytological change as well as discrepancy in size between
the mature egg and brood cavity and changes in embryonic
yolk content all point to the existence of EEN. Even if the
increase in embryo size is small and nutrient transfer is negligible, the morphological evidence strongly supports the
inference that some exchange (more than just of gases and
water) occurs between the embryo and the parent.
Another factor to consider is that of water absorption as a
reason for embryo enlargement. In studies on vertebrates,
measurements of changes in dry mass are currently the main
indicator of EEN whereas volume and wet mass are not considered as reliable criteria (Blackburn 1994 ). One reason is
that developing embryos always increase in wet mass and
volume (due to water uptake), regardless of whether matrotrophy is present.
In contrast with vertebrates, embryonic size increase is
still widely used as evidence of matrotrophy in invertebrates
and lower chordates. Experiments with radiolabelling and
diet manipulation (Toolson 1985 ; Hoese and Janssen 1989 ;
Frick 1998 ) as well as ultrastructural studies (Domenici and
Gremigni 1977 ; Cable and Tinsley 1991 ; Schwartz and
Dimock 2001 ; Korneva 2005 ) are very rare. Instead many
authors have recorded and described anatomical changes in
both the parent and the offspring during incubation, considering
such changes as additional evidence for matrotrophy (e.g.
Hagan 1951 ; Mukai et al. 1987 ; Farley 2001 ; etc.). The small
3.3 Evolution of Matrotrophic Incubation in Cheilostomata