242
assumed to be similar (Saunders and Metaxas 2010 ). The
long development of the cyphonautes larva may be explained,
among other things, by irregular food supply and by the fact
that some of the acquired energy is spent on feeding and
locomotion. In contrast, species of the malacostegine genus
Conopeum appear to have relatively short-lived planktotrophic larvae with a lifespan of a few days only (see Cook
1962 ; Dudley 1973 ). Dudley suggested that there is a trend
towards “reduction” of planktotrophic larva in the malacostegine genera Membranipora , Electra and Conopeum . The
largest and longest-living cyphonautes larvae are formed in
Membranipora , and the smallest ones, with the shortest life,
in Conopeum . Since egg size in malacostegines (and broadcasting ctenostomes) is fairly similar (being normally less
than 100 μm diameter, see Tables 3.1 and 3.2 ), it is clearly
does not affect the duration of larval life.
Theoretically, the increase in the amount of nutrients
transferred to the oocyte by the parent organism should result
in a shorter duration of development. If the nutritional
reserves are suffi cient to cover all needs to reach a competent
state and pass through metamorphosis, then feeding is not
required, and the duration of the larval period can be shortened. Indeed, on average, bryozoan endotrophic larvae
(formed from macrolecithal eggs) develop faster than
cyphonautes larvae, but again this is not very strict.
The developmental period of non-feeding bryozoan larvae consists of the incubation period during which
embryogenesis takes place and a free-swimming period
until larval settlement. In the laboratory, the latter period in
most bryozoan species studied is several hours to 1 day.
Only in a few species can large larvae swim for up to 4–5
days (Cook 1985 ). For the entire larval developmental
period until metamorphosis, an extreme example comes
from the descriptions of Paltschikova-Ostroumowa ( 1926 )
and Braiko ( 1967 ), who reported that embryogenesis in the
brood chamber of the cheilostome Tendra zostericola takes
from 10 h to 2 days. After that, according to observations in
the laboratory, the larva spends from 6–8 h to 2 days in the
water column before settlement. Thus the period from oviposition to metamorphosis takes from 16 h to 4 days. It
should be stressed here that the diameter of oocytes in this
species is only 70 μm (Braiko 1967 ), which is comparable
to the size of oocytes in cheilostomes with planktotrophic
larvae. Thus, the egg size being similar, development in
Tendra occurs faster than even in those gymnolaemate
broadcasters whose larvae have the shortest life (about a
week presumed for Conopeum , see above). A similar situation occurs in the ctenostome brooder Triticella fl ava whose
larvae develop from oocytes 65 μm in diameter during
approximately 8 days (Ström 1969 ). In addition, the small
egg size in these two species shows that premetamorphic
development is energetically not very costly (see also Byrne
et al. 2003 ).
Further comparison is hampered because of the very large
range of larval-development time (1–8 weeks) in broadcasters that all have small eggs of about the same size. Another
obstacle is the scarcity of data on the duration of larval development. In general, most gymnolaemate brooders have
larger eggs than broadcasters and their lecithotrophic larvae
develop faster than the longest-living planktotrophic larvae
(10–14 days vs 1–2 months in Electra and Membranipora ).
At the same time, the duration of development in brooders is
comparable to or possibly longer than that in short-lived
cyphonautes larvae (in Conopeum ). For instance, nonfeeding larvae of the ctenostome Bowerbankia gracilis
develop from eggs 350 μm in diameter in 12–14 days (Reed
1988 , 1991 ). According to Nielsen ( 1981 ), larval development in Pacifi cincola insculpta (egg diameter 250 × 225 μm)
took about the same time, i.e. 11–15 days in the sea and 6–15
days in the laboratory. In Fenestrulina miramara (as measured from the illustration, egg diameter is 320 × 270 μm),
larval development took 10–14 and 10–13 days, respectively,
under the same conditions. Interestingly, Silén ( 1945 )
reported that development of the larva of the cheilostome
Callopora dumerilii from a much smaller oocyte (120 μm in
diameter) also took two weeks (under laboratory conditions).
Thus from comparing developmental time in brooders, one
can conclude that, (1) larvae from eggs of strongly differing
size can take the same time to develop, and (2) larvae from
larger eggs ( Pacifi cincola , Fenestrulina ) can develop faster
than larvae from smaller eggs ( Callopora ). The latter conclusion accords with the suggestion that a reduction in development time may be correlated with egg enlargement. However,
the situation can be opposite, too, since development takes
just 8 days in T . fl ava (egg diameter 65 μm) and 12–14 days
in B . gracilis (350 μm). Also, the wide variation in larval
development time in Pacifi cincola insculpta should be noted.
At the same time, in some cheilostome species the duration of development of endotrophic larvae is comparable
with that of long-lived cyphonautes larvae. For instance,
brooded larvae of Cryptosula pallasiana in Nova Scotia
were developing in the aquarium for approximately 30
days (Gordon 1977 ) (oocyte diameter 180 × 150 μm, pers.
obs.). It is unclear whether this time corresponds to the
duration of larval development in nature, however. A similar duration has been reported for larvae of the matrotrophic brooder Celleporella hyalina , which take 3–4 weeks
to develop in natural conditions in north Wales (Cancino
and Hughes 1988 ) (oocyte diameter about 80 μm) although
the developmental time can be shorter, just 12–14 days
(Hughes 1987 ). The same egg size (80 μm) is characteristic of the matrotrophic cheilostome Bugula foliolata (as
B . fl abellata ), whose larva develops over two weeks (see
Corrêa 1948 ), and it seems that extraembryonic nutrition
does not increase larval developmental time, at least on
some occasions.
3 Evolution of Reproductive Patterns in Cheilostomata
assumed to be similar (Saunders and Metaxas 2010 ). The
long development of the cyphonautes larva may be explained,
among other things, by irregular food supply and by the fact
that some of the acquired energy is spent on feeding and
locomotion. In contrast, species of the malacostegine genus
Conopeum appear to have relatively short-lived planktotrophic larvae with a lifespan of a few days only (see Cook
1962 ; Dudley 1973 ). Dudley suggested that there is a trend
towards “reduction” of planktotrophic larva in the malacostegine genera Membranipora , Electra and Conopeum . The
largest and longest-living cyphonautes larvae are formed in
Membranipora , and the smallest ones, with the shortest life,
in Conopeum . Since egg size in malacostegines (and broadcasting ctenostomes) is fairly similar (being normally less
than 100 μm diameter, see Tables 3.1 and 3.2 ), it is clearly
does not affect the duration of larval life.
Theoretically, the increase in the amount of nutrients
transferred to the oocyte by the parent organism should result
in a shorter duration of development. If the nutritional
reserves are suffi cient to cover all needs to reach a competent
state and pass through metamorphosis, then feeding is not
required, and the duration of the larval period can be shortened. Indeed, on average, bryozoan endotrophic larvae
(formed from macrolecithal eggs) develop faster than
cyphonautes larvae, but again this is not very strict.
The developmental period of non-feeding bryozoan larvae consists of the incubation period during which
embryogenesis takes place and a free-swimming period
until larval settlement. In the laboratory, the latter period in
most bryozoan species studied is several hours to 1 day.
Only in a few species can large larvae swim for up to 4–5
days (Cook 1985 ). For the entire larval developmental
period until metamorphosis, an extreme example comes
from the descriptions of Paltschikova-Ostroumowa ( 1926 )
and Braiko ( 1967 ), who reported that embryogenesis in the
brood chamber of the cheilostome Tendra zostericola takes
from 10 h to 2 days. After that, according to observations in
the laboratory, the larva spends from 6–8 h to 2 days in the
water column before settlement. Thus the period from oviposition to metamorphosis takes from 16 h to 4 days. It
should be stressed here that the diameter of oocytes in this
species is only 70 μm (Braiko 1967 ), which is comparable
to the size of oocytes in cheilostomes with planktotrophic
larvae. Thus, the egg size being similar, development in
Tendra occurs faster than even in those gymnolaemate
broadcasters whose larvae have the shortest life (about a
week presumed for Conopeum , see above). A similar situation occurs in the ctenostome brooder Triticella fl ava whose
larvae develop from oocytes 65 μm in diameter during
approximately 8 days (Ström 1969 ). In addition, the small
egg size in these two species shows that premetamorphic
development is energetically not very costly (see also Byrne
et al. 2003 ).
Further comparison is hampered because of the very large
range of larval-development time (1–8 weeks) in broadcasters that all have small eggs of about the same size. Another
obstacle is the scarcity of data on the duration of larval development. In general, most gymnolaemate brooders have
larger eggs than broadcasters and their lecithotrophic larvae
develop faster than the longest-living planktotrophic larvae
(10–14 days vs 1–2 months in Electra and Membranipora ).
At the same time, the duration of development in brooders is
comparable to or possibly longer than that in short-lived
cyphonautes larvae (in Conopeum ). For instance, nonfeeding larvae of the ctenostome Bowerbankia gracilis
develop from eggs 350 μm in diameter in 12–14 days (Reed
1988 , 1991 ). According to Nielsen ( 1981 ), larval development in Pacifi cincola insculpta (egg diameter 250 × 225 μm)
took about the same time, i.e. 11–15 days in the sea and 6–15
days in the laboratory. In Fenestrulina miramara (as measured from the illustration, egg diameter is 320 × 270 μm),
larval development took 10–14 and 10–13 days, respectively,
under the same conditions. Interestingly, Silén ( 1945 )
reported that development of the larva of the cheilostome
Callopora dumerilii from a much smaller oocyte (120 μm in
diameter) also took two weeks (under laboratory conditions).
Thus from comparing developmental time in brooders, one
can conclude that, (1) larvae from eggs of strongly differing
size can take the same time to develop, and (2) larvae from
larger eggs ( Pacifi cincola , Fenestrulina ) can develop faster
than larvae from smaller eggs ( Callopora ). The latter conclusion accords with the suggestion that a reduction in development time may be correlated with egg enlargement. However,
the situation can be opposite, too, since development takes
just 8 days in T . fl ava (egg diameter 65 μm) and 12–14 days
in B . gracilis (350 μm). Also, the wide variation in larval
development time in Pacifi cincola insculpta should be noted.
At the same time, in some cheilostome species the duration of development of endotrophic larvae is comparable
with that of long-lived cyphonautes larvae. For instance,
brooded larvae of Cryptosula pallasiana in Nova Scotia
were developing in the aquarium for approximately 30
days (Gordon 1977 ) (oocyte diameter 180 × 150 μm, pers.
obs.). It is unclear whether this time corresponds to the
duration of larval development in nature, however. A similar duration has been reported for larvae of the matrotrophic brooder Celleporella hyalina , which take 3–4 weeks
to develop in natural conditions in north Wales (Cancino
and Hughes 1988 ) (oocyte diameter about 80 μm) although
the developmental time can be shorter, just 12–14 days
(Hughes 1987 ). The same egg size (80 μm) is characteristic of the matrotrophic cheilostome Bugula foliolata (as
B . fl abellata ), whose larva develops over two weeks (see
Corrêa 1948 ), and it seems that extraembryonic nutrition
does not increase larval developmental time, at least on
some occasions.
3 Evolution of Reproductive Patterns in Cheilostomata