72
is oviparous but did not noticed whether she observed
mature reproducing colonies. Later work showed that the
ITO develops at the onset of oogenesis, and thus only in
mature colonies (Hageman 1981 ). Thus, it is possible that
Corrêa observed non-fertile colonies. The broadcasting
freshwater ctenostome Hislopia malayensis , with multiserial colonies, lacks an ITO (Wood, personal communication,
2009), and this may be associated with the small number of
zooids in small colonies. In this case, there should be a high
rate of water exchange in a colony, quickly exiting eggs and
preventing them from being swallowed. It is also possible
that some broadcasters that secondarily acquired uniserial
growth may have inherited the ITO from their multiserial
broadcasting ancestors.
1.3.9.5 Secondary Loss of the ITO
It is feasible that the ITO could be lost secondarily owing to
(1) secondary acquisition of uniserial budding in broadcasters,
and/or (2) the evolution of brooding.
Secondary loss of the intertentacular organ (ITO) in multiserial brooders might have occurred because oocytes no
longer had to be transported away from the parent colony. In
gymnolaemates, eggs are incubated either on the zooid
surface or inside specialized brood chambers (Ostrovsky
2008a , c ; see also Chap. 2 ). The ITO theoretically could have
been present in early cheilostome brooders (Calloporidae)
with multiserial colonies and ovicells constructed of spines
(Ostrovsky and Taylor 2004 , 2005 ). With the assistance of
the ITO, mature eggs could pass directly to the brood cavity.
Such activity has been recorded in the cheilostome Tendra
zostericola where both the ITO and the tentacle crown enter
the cavity of the acanthostegal brood chamber where embryo
incubation takes place (Paltschikowa-Ostroumowa 1926 ;
Braiko 1967 ).
However, hypothesized oocyte enlargement during the
evolutionary transition to a lecithotrophic larva in brooders
(discussed in Chap. 3 ; see also Taylor 1988 ; Ostrovsky 2009 )
could make oviposition via the SNP more effective, and the
ITO might be lost. On the other hand, large oocyte size is
obviously not an obstacle in some instances, since they are
very fl exible in gymnolaemates. They squeeze not only
through a tiny supraneural coelomopore (Gerwerzhagen
1913 ; Silén 1945 ), but also through a tube-like ovipositor in
Schizoporella (Maturo 1991b ; Zimmer, personal communication, 2009). Large oocytes and a large ITO are also
described in the ovicell brooder Thalamoporella evelinae
(see Marcus 1941a ). Additionally, it should be mentioned
that secondarily uniserial brooders could have inherited a
SNP from their multiserial brooding ancestors.
Evidence from the literature shows that four brooding
species possess the ITO either in its complete or reduced
( Bulbella abscondita ) form (see above). Why should this be
the case? In B . abscondita the ITO has a role where it specifi -
cally manipulates the eggs, attaching them to the introvert
(Braem 1951 ). There is no specifi c activity of the ITO mentioned in the case of Alcyonidium duplex , in which released
eggs stick to the polypide diaphragm region (see Prouho
1892 ). In Tendra zostericola the ITO enters the large brood
chamber during oviposition and it is possible that similar
behaviour occurs in Thalamoporella evelinae . It should be
noted that several embryos are simultaneously incubated in
all four species mentioned, which is rare among gymnolaemates (Ostrovsky et al. 2008 ).
Based on the above considerations and the pattern of distribution of the ITO throughout the Gymnolaemata (see
Table 1.9 ), Ostrovsky and Porter ( 2011 ) theorized that the
ITO has been lost independently in congeneric species of
Alcyonidium , Victorella , Thalamoporella and perhaps
Bifl ustra , i.e. in both gymnolaemate orders; all of these genera include species with or without an ITO (see also Reed
1991 ). Most Alcyonidium (Ctenostomata) species are broadcasters with an ITO. The rest are introvert brooders with a
SNP, and only A . duplex has an ITO and a mixed type of
brooding, possibly representing the transitional stage from
broadcasting to internal brooding. All of them (except, to
some extent, A . albidum ) form multiserial colonies.
It was suggested that the brood chambers of Tendra and
Thalamoporella evolved independently of conventional
cheilostome ovicells (Harmer 1926 ; Ostrovsky and Taylor
2005 ; see also Chap. 2 ). The presence of an ITO in Tendra
zostericola and Thalamoporella evelinae supports this. Both
tendrids and thalamoporellids could have inherited the ITO
from broadcasting malacostegan ancestors, but it was later lost
in some species (e.g. in Thalamoporella prominens , which
possesses a SNP; see Marcus 1938a ). In contrast, in the cheilostome genus Schizoporella , the ovipositor may be a secondary novelty that evolved de novo, since it is positioned some
distance from the normal site of a supraneural pore (Zimmer
and Temkin, personal communications, 2009). Information
given by Reed ( 1991 ) about oviposition via the genital pore in
the ooecial vesicle in an unidentifi ed Schizoporella species
actually describes the ovipositor (Zimmer, personal communication, 2009). Cook ( 1985 , p. 49) recorded oviposition via the
coelomic pore between the two distal tentacles in S . fl oridana ,
however. Thus, as with Alcyonidium and Thalamoporella ,
different structures for oviposition (SNP and ITO) can be
present within the same genus.
On the other hand, some primitive cheilostome lineages
probably never possessed an ITO. For instance, some uni-/biserial cheilostome erect brooders (Eucrateidae, Leiosalpingidae,
Scrupariidae, Alysidiidae) could have evolved independently
from uniserial malacostegans with a supraneural coelomopore,
and Aeteidae from a uniserial ctenostome ancestor (Jebram
1992 ) [but the latter idea is not supported by Waeschenbach
et al. ( 2012 ), who grouped Aetea with malacostegines based on
molecular sequencing]. The independent origin of these
1 Reproductive Patterns of Gymnolaemate Bryozoa: General Overview and Comparative Analysis
is oviparous but did not noticed whether she observed
mature reproducing colonies. Later work showed that the
ITO develops at the onset of oogenesis, and thus only in
mature colonies (Hageman 1981 ). Thus, it is possible that
Corrêa observed non-fertile colonies. The broadcasting
freshwater ctenostome Hislopia malayensis , with multiserial colonies, lacks an ITO (Wood, personal communication,
2009), and this may be associated with the small number of
zooids in small colonies. In this case, there should be a high
rate of water exchange in a colony, quickly exiting eggs and
preventing them from being swallowed. It is also possible
that some broadcasters that secondarily acquired uniserial
growth may have inherited the ITO from their multiserial
broadcasting ancestors.
1.3.9.5 Secondary Loss of the ITO
It is feasible that the ITO could be lost secondarily owing to
(1) secondary acquisition of uniserial budding in broadcasters,
and/or (2) the evolution of brooding.
Secondary loss of the intertentacular organ (ITO) in multiserial brooders might have occurred because oocytes no
longer had to be transported away from the parent colony. In
gymnolaemates, eggs are incubated either on the zooid
surface or inside specialized brood chambers (Ostrovsky
2008a , c ; see also Chap. 2 ). The ITO theoretically could have
been present in early cheilostome brooders (Calloporidae)
with multiserial colonies and ovicells constructed of spines
(Ostrovsky and Taylor 2004 , 2005 ). With the assistance of
the ITO, mature eggs could pass directly to the brood cavity.
Such activity has been recorded in the cheilostome Tendra
zostericola where both the ITO and the tentacle crown enter
the cavity of the acanthostegal brood chamber where embryo
incubation takes place (Paltschikowa-Ostroumowa 1926 ;
Braiko 1967 ).
However, hypothesized oocyte enlargement during the
evolutionary transition to a lecithotrophic larva in brooders
(discussed in Chap. 3 ; see also Taylor 1988 ; Ostrovsky 2009 )
could make oviposition via the SNP more effective, and the
ITO might be lost. On the other hand, large oocyte size is
obviously not an obstacle in some instances, since they are
very fl exible in gymnolaemates. They squeeze not only
through a tiny supraneural coelomopore (Gerwerzhagen
1913 ; Silén 1945 ), but also through a tube-like ovipositor in
Schizoporella (Maturo 1991b ; Zimmer, personal communication, 2009). Large oocytes and a large ITO are also
described in the ovicell brooder Thalamoporella evelinae
(see Marcus 1941a ). Additionally, it should be mentioned
that secondarily uniserial brooders could have inherited a
SNP from their multiserial brooding ancestors.
Evidence from the literature shows that four brooding
species possess the ITO either in its complete or reduced
( Bulbella abscondita ) form (see above). Why should this be
the case? In B . abscondita the ITO has a role where it specifi -
cally manipulates the eggs, attaching them to the introvert
(Braem 1951 ). There is no specifi c activity of the ITO mentioned in the case of Alcyonidium duplex , in which released
eggs stick to the polypide diaphragm region (see Prouho
1892 ). In Tendra zostericola the ITO enters the large brood
chamber during oviposition and it is possible that similar
behaviour occurs in Thalamoporella evelinae . It should be
noted that several embryos are simultaneously incubated in
all four species mentioned, which is rare among gymnolaemates (Ostrovsky et al. 2008 ).
Based on the above considerations and the pattern of distribution of the ITO throughout the Gymnolaemata (see
Table 1.9 ), Ostrovsky and Porter ( 2011 ) theorized that the
ITO has been lost independently in congeneric species of
Alcyonidium , Victorella , Thalamoporella and perhaps
Bifl ustra , i.e. in both gymnolaemate orders; all of these genera include species with or without an ITO (see also Reed
1991 ). Most Alcyonidium (Ctenostomata) species are broadcasters with an ITO. The rest are introvert brooders with a
SNP, and only A . duplex has an ITO and a mixed type of
brooding, possibly representing the transitional stage from
broadcasting to internal brooding. All of them (except, to
some extent, A . albidum ) form multiserial colonies.
It was suggested that the brood chambers of Tendra and
Thalamoporella evolved independently of conventional
cheilostome ovicells (Harmer 1926 ; Ostrovsky and Taylor
2005 ; see also Chap. 2 ). The presence of an ITO in Tendra
zostericola and Thalamoporella evelinae supports this. Both
tendrids and thalamoporellids could have inherited the ITO
from broadcasting malacostegan ancestors, but it was later lost
in some species (e.g. in Thalamoporella prominens , which
possesses a SNP; see Marcus 1938a ). In contrast, in the cheilostome genus Schizoporella , the ovipositor may be a secondary novelty that evolved de novo, since it is positioned some
distance from the normal site of a supraneural pore (Zimmer
and Temkin, personal communications, 2009). Information
given by Reed ( 1991 ) about oviposition via the genital pore in
the ooecial vesicle in an unidentifi ed Schizoporella species
actually describes the ovipositor (Zimmer, personal communication, 2009). Cook ( 1985 , p. 49) recorded oviposition via the
coelomic pore between the two distal tentacles in S . fl oridana ,
however. Thus, as with Alcyonidium and Thalamoporella ,
different structures for oviposition (SNP and ITO) can be
present within the same genus.
On the other hand, some primitive cheilostome lineages
probably never possessed an ITO. For instance, some uni-/biserial cheilostome erect brooders (Eucrateidae, Leiosalpingidae,
Scrupariidae, Alysidiidae) could have evolved independently
from uniserial malacostegans with a supraneural coelomopore,
and Aeteidae from a uniserial ctenostome ancestor (Jebram
1992 ) [but the latter idea is not supported by Waeschenbach
et al. ( 2012 ), who grouped Aetea with malacostegines based on
molecular sequencing]. The independent origin of these
1 Reproductive Patterns of Gymnolaemate Bryozoa: General Overview and Comparative Analysis