Life cycle of the marbled octopus , Amphioctopus aegina ( Gray ) ( Cephalopoda : Octopodidae ) reared in the laboratory

Reproducing small eggs and planktonic hatchlings is a reproductive strategy of many species of benthic octopods although it is considered a pleisiomorphic state. The young in the planktonic and settling phases have a high energy consumption and require live food of specific size range characteristics, causing difficulties in obtaining appropriate food organisms for rearing experiments through the entire life cycle. This study obtained information on details of life cycle characteristics of Amphioctopus aegina (Gray) by resolving these difficulties. The aspects of life cycle of A. aegina were similar to those of other benthic octopus with a similar reproductive strategy. Growth was allometric, consisting of 3 phases in terms of body weight, including a transitional phase for the early settling stage. Feeding and conversion efficiency peaked over a 2-month period prior to reproduction, representing the period of energy storage. Growth from hatching to spawning took approximately 74% of the entire life span and the reproductive phase itself took 34%. Such longevities are similar to those of other benthic cephalopods with planktonic hatchlings. Overall similarities in these cephalopod taxa attest to the fitness of a life-history strategy involving production of planktonic offspring from benthic adults. A. aegina is the second benthic octopus species with planktonic hatchlings for which the life cycle has been completed through rearing in the laboratory.


INTRODUCTION
One major aspect of the evolutionary adaptations of benthic octopods is exhibited in 2 life-history strategies.The first one is production of relatively few large eggs resulting in larger hatchlings with a benthic mode of life.The second one is production of numerous small eggs resulting in hatchlings with a planktonic mode of life.Boletzky (1974Boletzky ( , 1977Boletzky ( , 1978Boletzky ( -1979) ) proposed that when egg length is more than 10% of mantle length of adults, the resulting hatchlings are benthic (with relative mantle length of 6-20%) while egg lengths of less than 10% of adult mantle length are associated with planktonic hatchlings (with a mantle length of 2-5%).Producing large eggs, and hence benthic hatchlings, is considered an apomorphic state and producing small eggs and planktonic young is a pleisiomorphic state (Boletzky, 1977).The disadvantages of the planktonic mode of life for young octopods are higher energetic cost and higher risk of predation compared with the benthic mode.However, in the context of the evolutionary trade-offs, the advantage of the planktonic mode is that the planktonic hatchlings possess considerable potential for dispersal, assuring the distribution of the species (Villanueva and Norman, 2008).The planktonic life stage of the young octopods also enables them to occupy different niches and potentially to locate alternative food sources, allowing higher rates of energy consumption.These abilities could explain why a pleisiomorphic state such as the planktonic mode of life is still maintained through natural selection in species with wider distribution.
Information on life cycle characteristics, including feeding, growth and behaviour, can be obtained through captive rearing.The life histories of cephalopods are highly variable, deriving from the plasticity of their growth as affected by various biotic and abiotic factors.Forsythe and Van Heukelem (1987) observed that information on cephalopod growth can be derived from both laboratory rearing and field surveys.The major disadvantage of captive rearing is that it is unnatural or artificial.However, the advantage of rearing is that growth and life cycles can be measured and monitored under known conditions; environmental factors can be controlled or monitored and the impact of predation, density and food abundance is removed or controlled.Indeed, most of the available information on octopod growth over their life cycles comes from laboratory rearing (Hanlon, 1987) and, as mentioned above, concerns species with benthic hatchlings.At present, 24 species of octopods with planktonic hatchlings have been reared but hatchlings of only 2 species have been reported to develop to the settling stage: E. dofleini and O. vulgaris (Nabhitabhata, 1985(Nabhitabhata, , 1995;;Nabhitabhata et al., 2003;Villanueva and Norman, 2008).Only O. vulgaris has been reared through its complete life cycle in captivity (Iglesias et al., 2004).Other benthic cephalopods with planktonic hatchlings and benthic adults, Sepiella inermis Ferrussac and d'Orbigny (Nabhitabhata, 1997) and Euprymna hyllebergi Nateewathana (Nabhitabhata et al., 2005), could also be reared through their complete life cycles.Finding whether those taxa have common characters of life cycles is informative, inferring the evolutionary fitness of this life-history strategy.
Amphioctopus aegina (Gray) is a moderate-sized benthic octopus with a mantle length 40-100 mm.Fresh specimens can be recognized by the coloured longitudinal stripe on their dorsum (Fig. 1) with a cream transverse bar on the dorsal head between the eyes (Nateewathana, 1997;Norman, 1998Norman, , 2000)).The colour of the dorsal stripe varies among localities and Norman (2000) has suggested that A. aegina is a species complex.Distribution is wide in the coastal zone of the Indian and Western Pacific Ocean, from 30° to140°E and 40°N down to 20°S (Roper et al., 1984).In Thai waters, this species is common, living in muddy areas (Nateewathana, 1997).Norman (1998).Phanichpong (1985) described the reproductive biology and behaviour of A. aegina (as O. dollfusi) and suggested that this species reproduced all year round in Thailand, with the highest peak from March to May (summer to the early rainy season).The ovary weight of mature females was normally 5-7% of the body weight, reaching a maximum of 17% during the peak of reproduction.The testis weight of mature males was approximately 3-4% of body weight (with a maximum of 9.4%) throughout the year (Phanichpong, 1985).Mating, following the "enveloping pattern" (Gabe, 1975;Hartwick, 1983), was also observed in the laboratory.Other aspects of the life cycle of A. aegina remain little known, though it is an economically important species in Thailand.About 15000 t of octopus are fished annually from Thai waters, about 12% of the total cephalopod catch.This includes A. aegina (Supongpan, 1995), which is treated under the name Octopus dollfusi Robson, its junior synonym (Nateewathana, 1997;Norman and Hochberg, 2005a).
The outstanding problem for laboratory rearing of octopods is that the planktonic young octopods require live food of a specific size range and characteristics (Hanlon, 1987;Boyle and Rodhouse, 2005).The subsequent transitional settlement period also involves a transition to utilizing a new type and size of food organisms.The present study aims to resolve these difficulties for A aegina, making it only the second benthic octopus species with planktonic hatchlings (and the first such species of moderate size) for which the life cycle has been completed, and to determine details of morphology, feeding, growth and behaviour of A. aegina throughout its life cycle, including its embryonic phase and post-hatching to adult development stages.
Information on the life cycle and growth of A. aegina is potentially valuable to underpin sustainable fishing of the species and to provide a better under-standing of, and more detailed insight into, how egg size and juvenile habitat in octopods relate to other life cycle features, compared with other benthic cephalopods with which their niches overlap, i.e. those octopus species with a similar reproductive strategy and also other benthic cephalopods with planktonic offspring.In addition, determination of the morphological characters of the early life stages of this species will help to clarify previous taxonomic identifications based on wild planktonic young without a direct link to the spawning adults.

Maintenance and rearing of live specimens
Broodstock of A. aegina was collected live from octopus traps operating along the coast of Rayong Province in the eastern part of the Gulf of Thailand, South China Sea, Pacific Ocean.Twelve females with an average dorsal mantle length (ML) of 60.5±4.28 mm and a weight of 78.39±9.65 g and 10 males of 57.00±4.54mm and 49.00±10.19g were maintained on-board and then transported to the cephalopod hatchery of the Rayong Coastal Fisheries Research and Development Centre.The octopus were separated by sex and reared for 7 d in 1800-mm-diameter circular concrete tanks with a capacity of 2 m 3 and 300 mm water depth.Pieces of PVC pipe (50 mm diameter, 150 mm length) were supplied as shelters in the tanks.The tanks were equipped with aeration devices and supplied with running filtered seawater in order to minimize any temperature change.The octopus were fed twice daily at 08:00 h and 17:00 h.Tanks were cleaned by siphoning out the water and about 50% by volume was replaced daily.One pair of octopuses was then reared in each tank under similar conditions.The octopuses mated and spawned in the tanks.After spawning, the male was excluded.Hatchlings were reared in 50-L-capacity cylindrical PVC tanks with dark blue tank walls, a water depth of 150 mm and a stocking density of 1 individual•L -1 .The tanks were equipped with air stones attached to half-circular PVC pipes in order to direct the current.Hatchlings were fed ad libitum twice daily with live prawn larvae in a zoeal stage, Macrobrachium sintangense (De Man) (available from commercial hatcheries) of 1-2 mm total length for 30 d after hatching (Fig. 2).Eighteen batches of the octopus were studied from May 2006 to April 2007.
Live amphipods, Unciolella sp.Chevreux, wild palaemonid shrimps, Palaemon styliferus (H.Milne-Edwards) and dwarf gobies, Stigmatogobius romeri (M.Weber), collected from the wild, were fed to the young octopuses at their early settling stage for 20 d (days 30-50).The water depth was gradually increased to 300 mm at 30 days.Tanks were cleaned by siphoning out the water and about 80% by volume was replaced daily.The settling young were transferred to 1800-mm-diameter circular concrete tanks with grey tank walls and maintained at a density of 1 individual.L -1 .Water depth was decreased to 150 mm and coral pebbles were supplied as shelters.After 50 d, the octopuses were trained to feed on dead fish meat, Selaroides leptolepis (Cuvier).Live bivalves, Donax spp.Linnaeus, Anadara granosa (Linnaeus) and Perna viridis (Linnaeus) and prawns (M.santangense) were also fed to the octopuses.Pieces of PVC pipe (50 mm diameter, 150 mm length) were supplied as shelters.The survival of the octopuses was not recorded.Other details of tank management followed Nabhitabhata et al. (2005).The average temperature during the experiment was 29.7±0.6°C,pH was 8.1±0.1 and salinity was 30.0±1.3 psu.The photoperiod in the tanks was natural throughout the study period.

Behaviour
Aspects of octopus behaviour, such as details of life history and as information for further feed and tank management, were observed daily in live specimens.Recordings were taken in the form of drawings, still photographs and videos.The octopuses in all tanks were observed during the planktonic phase.After settlement, 175 octopus were randomly selected every 3 h daily and observed for a period of 2 h.

Feeding and growth
The numbers of individuals of live food organisms were recorded daily before feeding and their weight was calculated from the average figures previously obtained from a sample of 50 individuals.Only weight was recorded for the dead food.
The feeding rate and gross food conversion efficiency were determined every 10 days (N=153 individuals) and calculated according to Choe (1966a, b): where FR is feeding rate (% d -1 ), GFCE is gross food conversion efficiency (%), F is total food consumed on a wet weight basis (g), W is average weight (g) during time period t, W 1 is initial weight (g), W 2 is final weight (g), and t is number of days (10-day period).
Growth was determined every 10 d in terms of the gain in dorsal mantle length (ML, mm) and wet body weight (W, g) (N=153).Ten octopuses were sampled from each tank and their activity during measurement and weighing was reduced by decreasing the water depth in the container.The instantaneous rates of relative growth (IGR) in both ML (IGRL, %) and weight (IGRW, %) were calculated according to Forsythe (1984) and Forsythe and Van Heukelem (1987): where ML 1 is the initial mantle length (mm) and ML 2 is the final mantle length (mm).
The body weight-mantle length was expressed using a power regression model (N=153).The mantle length-age relationship was expressed by an exponential regression model in the early growth phase (N=47) and a cubic regression model in the following phase (N=87).Three different types of model were used to fit the weight-age relationship, exponential regression (N=38), power regression (N=40) and quadratic regression (N=63).
where a 1 , a 2 , …a n and b are constants and A is the age (d).

Eggs
The egg capsule is single, with a rice-grain shape and a fine stalk attaching it to the festoon (Fig. 3).Clusters of eggs are attached together forming a strand of eggs.The size of each egg capsule was 3.17±0.06mm (N=360) along its major axis and 1.25±0.04mm along its minor axis and the weight was 0.19±0.01mg.The newly laid capsule is opaque white and after 5 d it gradually turns more transparent until the embryo becomes visible.
The egg is telolecithal.Eight-cell cleavage was observed at 13 h after fertilization.Organogenesis occurred from 129 h (day 5).The first inversion of the embryo was observed from 137 h (day 5), turning the posterior mantle towards the stalk of the capsule in head-up position.The red colour of the retina was observed from 190 h (day 7).Chromatophores were observed from 227 h and the black retina and an ink sac from 238 h (day 9).The second reversion was from 335 h (day 13), when the embryo inverted to its original position with its head towards the stalk of the capsule, corresponding to the hatching position.The first hatching occurred at 385 h (day 16) (Fig. 4).
The embryonic phase had an average duration of 18.3±0.6d (range 16-22 d, N=20) at 29.7±0.6°C.Hatching mostly occurred at night.The hatching period of the eggs in the same cluster of a single female continued for a period of 3-8 d from the first to the last egg.

Behaviour
The hatchling was planktonic, hovering headdown with its mantle at a 45° angle to the tank bottom.Feeding was observed immediately after hatching.The hovering position was also the ambushing position of the hatchlings for a dart-motion to attack passing prey.Their colour pattern turned from transparent to a darker one during the hunting mode of behaviour.Prey was seized using the arms, which, in hatchlings, were each equipped with 7 suckers.The internal yolk sac was capable of sustaining unfed hatchlings for at least 5 d.
The planktonic phase lasted 20-25 d and the octopuses subsequently entered their settling stage.From day 15, the early settlers adopted the habit of temporarily attaching themselves to the tank wall and gradually, with a longer period of attachment, increased the depth of attachment down to the tank bottom.During the transitional settlement period, the depth of attachment was gradually increased by 50 mm/d, approximately.The young still ambushed their prey in the water column during this transitional period.The last group of settlers entered the benthic stage and fed on benthic prey after 30 d. Crepuscular behaviour was exhibited after    settlement, with small individuals living under shelters and larger individuals living in holes or 'homes'; all animals were solitary.Cannibalism was occasionally observed in the rearing tanks from hatching to about 70 d.The size difference between predator and prey was always more than 50%.

Feeding and growth
Hatchlings of the marbled octopus grew from an average mantle length of 2.68±0.09mm (ML) and an average weight of 0.003±0.001g (N=500) to a ML of 6.34±0.91 mm ML and a weight of 0.068±0.027g (N=10) in 30 d (Figs. 5, 6).The instantaneous relative growth rate (IGR) during this period was 4.14% by length and 12.90% by weight.The IGR was highest between 30 and 40 d (i.e. after settlement), when it reached 5.06% by ML (Fig. 5) and 17.45% by weight (Fig 6).After that, IGR rapidly decreased to achieve relative stability after 80 days, when maturity was reached because the male hectocotylus was recognized.The average IGR from hatching to 140 days of age was 1.94±1.78% in ML and 6.41±5.60% in weight (N=14).
Food consumption of the octopus averaged 53.60±67.13%body weight.d - (N=14).The feeding rate peaked at 239.9% after 30-40 d, at settlement, and then rapidly decreased to 12.8% at 80-90 d, as the animals entered maturity (Fig. 7).In contrast to the feeding rate, the food conversion efficiency decreased from 6.2-10.58% in the planktonic stage to 5.86% at settlement.The peak efficiency was 31.9-34.6%at 50-70 d, when the animals were entering maturity (Fig. 8), and then decreased to 10.5% at 140 d of age.The average efficiency from hatching to 140 d was16.04±10.98%(range 5.86-34.59,N=14).
The relationship between the mantle length (mm) and weight (g) from hatching to 210 d could be expressed as a power regression (Fig. 9):    Pair formation prior to mating was not observed.The mating behavioural sequence followed 1 of 2 patterns.The first one was by tactile contact or by an enveloping pattern (Fig. 12a).The male displayed a dark brown colour pattern on half of his mantle, approached the female in her home, then mounted and spread his arm-web umbrella to envelop the female.The female covered her own mantle with her arm-web umbrella, facing (suckers outward) towards the male.The male displayed a dark brown colour and then probed his hectocotylus into the female mantle cavity.Copulation took 7-10 min.After that, the female pushed the male out of her home with her arms.If the female rejected mating, she attacked the approaching male with a rapid forward motion from her home with an attempt to envelop the male.
The second pattern was remote copulation (Fig. 12b).The male approached the female in her home, extended his hectocotylus and inserted it into the mantle cavity of the female.Copulation was without any other contact of other parts of body from both sexes.This pattern of copulation took 3-5 min.Other behaviour observed before and after copulation was similar to the enveloping pattern.
The interval between mating and spawning averaged 32.3±12.9d (range 20-40 d).On the day of spawning, the female did not leave her home for feeding.An egg cluster was cradled in her arm web behind her mantle.The female regularly cleaned and aerated her egg clusters with her arms and directed jets of water from her funnel.Some brooding females initiated premature hatching of her eggs by a strong and rapid exhalent jetting and shaking of egg clusters with her arms after being disturbed by sudden visual stimuli (i.e.approach of tank-cleaning devices).Cessation of food intake by the brooding female was observed until hatching and, in most cases, until her death.Some females fed for 1 to 2 d after hatching before ceasing.The average number of eggs laid by a single female octopus was 6900.0±1417.7 eggs (N=20).
tropical-moderate size reveals that there are similarities in many aspects of their life cycle irrespective of either their habitats or size at maturity.The planktonic period, hatching to settlement, was approximately 30-40 d in most of the species and a week shorter in A. aegina in this study (Table 1).A lower temperature (20°C) prolonged the period in larger species, O. vulgaris (Villanueva, 1995) and Enteroctopus dofleini (Hartwick, 1983).The shortest period, 14 d, was recorded for O. rex at 27.5°C (Nabhitabhata et al., 2003).Although it was comparable to 20 d at 29.7°C for A. aegina in this study, it might also be species-specific.However, the similar size at settlement of approximately 4-10 mm ML (Table 1) in all species suggests that the duration of the period before settlement is less important than the actual size at settlement.The daily feeding rates in temperate octopods were less than 10% at 20-25°C compared with the approximately 53.6% of A. aegina in this study.The lower feeding rates yielded higher conversion efficiencies, of approximately 60-75% in temperate compared with 16.0% in tropical A. aegina.However, the instantaneous daily growth rate, by weight, of A. aegina had a maximum value of 17.5%, much higher than the values of less than 10% for other species (Table 1).
The food conversion efficiency was lower in the planktonic phase than in the following phase, showing an opposite trend to the difference in feeding rate (Fig. 7, 8) and revealing a high energy cost for the planktonic mode of the paralarvae.The conversion efficiency peaked from the age of 60-70 d to 120-130 d, a period of 65 d before mating was observed at the age of 125 d.This two-month period indicates that the process of energy storage for subsequent reproduction required a relatively long period in A. aegina, because it takes as long as 27.9%, approximately one-third, of its lifespan, which is similar to the value of 34% (73 d) for the reproductive phase itself (Table 2, 3).
The daily growth rate was notably similar among the different benthic cephalopods, at approximately 2% in terms of mantle length and 3-7% in weight (Table 2).The present study supports the idea that the growth pattern could not be fitted by one single model throughout the entire life cycle because a single model will not properly describe the growth during early or late life (Forsythe and Van Heukelem, 1987).Such patterns were also reported in other tropical benthic cephalopods, e.g. S. inermis (Nabhitabhata, 2002) and E. hyllebergi (Nabhitabhata et al., 2005).These species showed 2 phases of growth during the life cycle (expressed by 2 models), when considered in terms of either mantle length or body weight.The present study also supports the statement by Forsythe and Van Heukelem (1987) that the description of growth at the intersection of the 2 equations is often not accurate.We found that adding a third model for the early settling phase of A. aegina could yield a better fit overall for the growth pattern over the course of the life cycle.However, the precise form of the models estimated Table 4. -Benthic octopods, inhabiting Thai waters, known to produce planktonic hatchlings (H, from rearing experiment) or assumed to do so due to producing small-type eggs (E, according to Boletzky 1974Boletzky , 1977Boletzky , 1978Boletzky -1979)).Norman and Hochberg (2005b) in this study might be species-specific and applicable only for octopuses in captivity, since the growth and life span of the octopus may be modified in captivity.Outliers in the models obtained are genuine points and not mistakes, which might be because of mutations or other unknown conditions in aquaculture.Those octopuses could survive in rearing tanks at least to the time of the latest growth measurement, but might not be able to do so in the wild.One notable characteristic is the egg brooding behaviour, which involved cradling egg strings on the arm web of the female octopods in Thai waters.This characteristic has been observed in reared O. membranaceus Quoy and Gaimard (Nabhitabhata, 1985), A. neglectus (Nateewathana and Norman) and O. cf.kaurna Stranks (Nabhitabhata, unpublished data), A. rex (Nateewathana and Norman) (Nabhitabhata et al., 2003), A. aegina (present study), Cistopus indicus (Rapp), and H. cf.maculosa (Nabhitabhata, 1995).The above-mentioned species all produce planktonic hatchlings.Egg cradling seems to be a common characteristic among octopods in Thai waters but it contrasts with what is seen in most species of octopods with planktonic hatchlings, which attach their egg strings to hard substrates (Villanueva and Norman, 2008).It is possible that the absence of preferred substrates in captivity might cause the octopus to cradle their egg strings instead of attaching them.Rocha et al., (2001) proposed simultaneous terminal spawning as one of the cephalopod reproductive strategies.The representative species for this strategy include 3 benthic octopods, O. cyanea, O. mimus and O. vulgaris.Although the reproductive behaviour and growth observed in this study indicates that spawning of A. aegina is terminal, further investigation on the oocyte maturation is required to fully characterize its reproductive strategy.
Thai waters, particularly in the Gulf of Thailand, are shallow (maximum depth 90 m with an average depth of about 30 m) with a fluctuating tropical environment (a long rainy season of at least 6 m and, consequently, a heavy freshwater discharge).A notable feature is that 17 of the 19 species of benthic octopods recorded from Thai waters (Nabhitabhata, 1999) are known or tend to reproduce planktonic hatchlings (Table 4).Rocha et al., (2001) also proposed that the strategy of simultaneous terminal spawning is adapted to extremely unstable environments.The 3 representative octopods of this strategy are known to reproduce planktonic hatchlings (Table 1).Two of them, O. cyanea and O. vulgaris, are also recorded in Thai waters (Table 4).This suggests that reproducing planktonic hatchlings is the reproductive strategy relevant to simultaneous terminal spawning in a neritic-fluctuating or unstable environment of benthic octopods.
The lifespan of the tropical A. aegina in this study, approximately 7 m (5-8 m), is short compared with those of the larger temperate species (Table 1).However, when compared wth other medium-size tropical benthic cephalopods (Table 2), it is around 0.4-1.2times as long.Although the period from hatching to complete settlement of A. aegina, is comparable to that of E. hyllebergi, after that the period to reach maturity is about twice as long.The neritic-benthic cephalopods also display consistent features in their reproductive phase.The longevity from hatching to maturity is about 60% to mating and 70-80% to spawning and the reproductive phase is about 30%, one-third of the lifespan (Table 3).
Overall, the life cycle of A. aegina is similar to that of other benthic octopods, with similar reproductive strategies in terms of its characteristics and habits in the embryonic, planktonic and settling phases.However, the growth pattern observed, which includes a single phase during transitional settlement, was determined under captive rearing conditions and might be modified from what occurs in the wild.The similarity of the duration of different phases of the life span of A. aegina compared with other benthic cephalopods with planktonic hatchlings indicates that they have adapted to a similar way of life although their niches overlap.The similarities among these benthic cephalopod taxa attest to the fitness of their life-history strategy of reproducing planktonic offspring.

Fig. 2 .
Fig. 2. -Diagram of feeding of Amphioctopus aegina on various live (full line) and non-live (dashed line) food corresponding to the age (d) after hatching.

Fig. 5 .
Fig. 5. -Growth in terms of mantle length (plotted lines, ML: mm), the instantaneous relative growth rate (continuous lines, IGRL: %) and age (d) after hatching.The horizontal line with figure indicates the average value for IGRL.Arrows indicate settlement (st), mating (m) and spawning (sp).

Fig. 6 .
Fig. 6. -Growth in terms of weight (plotted lines, W: g), instantaneous relative growth rate (continuous lines, IGRW: %) and age (d) after hatching.The horizontal line with figure indicates the average value for IGRW.Arrows indicate settlement (st), mating (m) and spawning (sp).

Fig. 7 .
Fig. 7. -Feeding rate (% weight) during the growth at age (d) after hatching.Arrows indicate settlement (st), mating (m) and spawning (sp).The horizontal dashed line with figure indicates the average value.

Table 1 .
-Comparison of the life cycle of octopus species known to produce planktonic hatchlings.

Table 2 .
-Comparison of the life cycle characteristics of reared benthic cephalopods in relation to age (d) and daily growth (%).

Table 3 .
-Comparison on the longevity of the stages of the life history of reared benthic cephalopods as a percentage of their lifespan.(Numbers in parentheses are estimated longevity when the embryonic phase is included in the life span).