Larval development of the mud shrimp Axianassa australis (Decapoda: Thalassinidea) under laboratory conditions

Ovigerous females of Axianassa australis were obtained from intertidal mudflats in south Texas, and eggs were hatched in the laboratory. Zoeal stages (Z) and the first postlarval stage (decapodid) were obtained, though survival rates were low. A few larvae passed through eight zoeal stages before moulting to the decapodid. Duration ranged between 3 and 4 days for each stage between ZI and ZVII, and 4 and 5 days for ZVIII, which totalled about 1 month for the entire zoeal phase of development. Three specimens that reached the decapodid stage were preserved for morphological descriptions so duration of this stage is unknown. Morphology of the first three zoeal stages and decapodid stage is described and compared with that of related thalassinideans reported in literature. As possible, comparisons are also made with larvae reared from conspecific populations in east Florida and Brazil. Biogeographic implications of the larval life history are discussed.


Introduction
Larval life history of the genus Axianassa is poorly known compared to that of other thalassinideans. The adult habitat (mud flats associated with mangroves or other vegetation) and the apparently long planktonic larval history complicate collection of ovigerous females and rearing of a full larval series. Larvae from wild plankton samples taken in the Gulf of Mexico have been previously referred to the genus Axianassa, though the species identification has remained in question (Ngoc-Ho 1981). At the time those larvae were described, no member of the genus Axianassa was known to occur in the southwestern Gulf of Mexico. However, the number of stages in plankton samples from the Gulf of Mexico suggested that they belonged to a species with a long water column life history and thus potential for broad dispersal. Subsequently, very similar zoeal stages were reported from waters off the coast of Louisiana (Truesdale and Andryszak 1983). While morphology of the ZI of Axianassa australis Rodriguez and Shimizu, 1992, was included with the species description of adult material from Brazil, apparent differences between the ZI of A. australis and the congeneric larvae described from the Gulf of Mexico suggested that these were not the same species (Table I).
In the present study, ovigerous females of A. australis were collected on the coasts of Texas and Florida. Larvae were subsequently hatched and reared in the laboratory. Table I. Comparison of three larval descriptions for genus Axianassa.

Rostrum
As long as antennae As long as antennae As long as antennae Eyes Round Morphology was examined for comparison to existing descriptions of both congeneric and other less closely related forms, as well as to recently obtained photographs of the ZI stage from conspecific Brazilian populations.

Materials and methods
Ovigerous females were collected from mud flats near Fort Pierce, Florida (Florida population) and Port Isabel, Texas (Texas population) with yabby pumps (as in Felder 1978). Adult females were maintained in 20-cm diameter finger bowls (at 27uC, salinity 35, psu scale) with daily water changes until eggs hatched. Larvae of the Texas population were reared in seawater taken offshore of Louisiana, filtered through steel wool, and aerated before use in cultures. Larvae of the Florida population were reared in water collected off Fort Desoto Park, Florida, otherwise treated as mentioned above.
To estimate stage durations, ZI larvae were moved to individual compartments of plastic trays upon hatching and maintained at 27uC, in filtered seawater (salinity 35, psu scale), on a 12 h light: 12 h dark cycle. Each day larvae were moved to clean trays with new seawater, fed freshly hatched nauplii of Artemia (Great Salt Lake), and examined to assess their stage of development. Observations were terminated when larvae expired or when the last few surviving decapodids were preserved. While most larvae died before reaching ZIV, two (out of 102) larvae from a single parental female of the Texas population were reared successfully through the decapodid stage. Larvae from other parental females of the Texas and Florida populations were also reared, but all died before the full course of zoeal development was completed.
Morphological studies were based on stages harvested from mass cultures and on moults from animals reared individually. Mass-cultured larvae were maintained in 20-cm diameter finger bowls (100-200 individuals per bowl) with daily water changes under conditions mentioned above. Animals were fixed in 70% ethanol, stained with chlorozol black, and transferred to glycerine prior to dissection. Unless otherwise noted, at least 10 animals were dissected for each described stage, and both right and left appendages were used for setal counts. Line illustrations were made on a Nikon inverted microscope fitted with a camera lucida. Measurements were made with a calibrated ocular micrometer. Carapace length (CL) was measured from the tip of the rostrum to the posterior midpoint of the carapace, and total length (TL) was measured from the rostral tip to the posterior midpoint of the telson. For each appendage, the arrangement of setae was listed sequentially from proximal to distal margins as in Nates et al. (1997) and Konishi (1989). Setal groups on successive segments were separated with a comma (,). Groups of setae on the same segment, or on different lobes of the same segment, article or endite, were separated with a plus (+). A question mark (?) was used to designate questionable distinctions between setae and aesthetascs. Roman numerals were used to describe the pattern of processes on the posterior margin of the telson.

Results
Prior to hatching, eggs changed gradually in colour from bright orange-red to a translucent brown. Late-stage embryos that were dropped by the female before hatching and larvae that hatched as prezoeas were not viable, even with aeration. Prezoeal stages did not appear to be fully developed and could not swim. Although small in size, zoeal stages fed readily on nauplii of Artemia.
Larvae of Axianassa australis passed through eight zoeal stages in laboratory cultures before moulting to the decapodid (Figure 1), but these data are based on only three animals that survived the full course of larval development. Duration of each stage from ZI to ZVIII ranged from 3 to 5 days each, and larvae thus required approximately 1 month to reach the decapodid stage. While the sample size is small (only five animals survived past ZIII and only three reached the decapodid stage), this estimate suggests a lengthy period for potential planktonic dispersal.
The zoeal stages were found to have distinctive chromatophore patterns ( Figure 1). Two large red chromatophores were present consistently on the telson and posterior margin of the last abdominal segment. One to three small, red chromatophores were present on the lateral margins of the carapace. Although not always easily observed at ZI, chromatophores on the carapace were typically well developed by stage ZIII. Although only two decapodids were available for dissection (one died prior to preservation), specimens did not vary in any marked way. Although setal numbers on various appendages may have varied, the ranges (between left and right sides) overlapped.
Because of high mortality in the moult to ZIV, the need to keep the survivors alive for studies of stage durations and the need to get at least some individuals through complete larval development, few materials were available for descriptions of the later zoeal stages. Thus illustrations and descriptions were limited to the first three zoeal stages (which most comparative materials are also limited to) and decapodid. Stages ZIV-ZVIII were, however, found to vary considerably in morphology from the few specimens observed.
Setal patterns are given from proximal to distal segments for each appendage. Where setation varies, illustrations represent the most common morphology observed. Carapace length (CL) and total length (TL) are given as the mean¡standard deviation (in mm), followed by the range of measurements from 10 specimens unless otherwise noted (as in stage D where only two specimens were available).
Antenna ( Figure 2b). Protopod with one distal spine between rami, one specimen with small distal spine near base of endopod; endopod with three long, plumose setae; scaphocerite (exopod) armed with one strong distolateral spine, 8-10 plumose setae on inner margin. Figure 2c, d). Mandibles asymmetrical; right mandible with two prominent teeth at base, lower plate with numerous small teeth, distal end pointed but not as strongly produced as on left mandible, sometimes with small denticles; left mandible sickle-shaped, distal end pointed with several teeth, inner surface of base with three to five teeth. Figure 2e). Coxal endite with three to four marginal setae; basal endite with two setae and two large dentate spines; endopodal lobe not distinct from protopod, with two to three setae on distal margin; protopod without setae. Maxilla (Figure 2f). Coxal endite bilobed, zero or one setae on proximal lobe, and three to four setae on distal lobe; basal endite bilobed with four to five setae on proximal lobe, three to five setae on distal lobe; endopod with one long and one short terminal setae; scaphognathite with three to five plumose setae.
Antenna (Figure 3b). Protopod with one well-developed distal spine between rami and one small distolateral spine near base of endopod; endopod with two to three long, plumose setae; scaphocerite (exopod) armed with one strong distolateral spine, 10-12 plumose setae on inner margin, one specimen with seta on scaphocerite spilt into two branches at distal end.
Maxillule (Figure 3e). Coxal endite with three to four marginal setae; basal endite with two to three setae and two to four large dentate spines on distal margin, a single specimen with one seta on lateral margin; endopodal lobe not distinct from protopod, three setae on distal margin; protopod without setae.
Antenna (Figure 4b). Protopod with one well-developed distal spine and one small distolateral spine near base of scaphocerite (exopod); endopod with one to three long, plumose setae; scaphocerite armed with one strong distolateral spine, 10-13 plumose setae on inner margin, one specimen with seta on scaphocerite split into two branches at distal end.
Maxillule (Figure 4e). Coxal endite with three to four marginal setae; basal endite with three setae and three to five large dentate spines on distal margin, a single specimen with one seta on lateral margin; endopodal lobe not distinct from protopod, three setae on distal margin; protopod without setae.
Maxilla (Figure 4f). Coxal endite bilobed, zero or one setae on proximal lobe, three to four setae on distal lobe; basal endite bilobed, four to six setae on proximal lobe, three to five setae on distal lobe; endopod with one long and one short terminal setae; scaphognathite with five to nine plumose setae, sometimes with row of setal hairs on mesial margin. Maxilliped 3 (Figure 4i). Two segments present, distal segment with six plumose setae.
Pereopod 2 (Figure 4k). Endopod produced as small lobe without setae, exopod differentiated from basis, without setae in most specimens; one specimen with developed appendage bearing four to five setae on exopod.
Pereopod 3 (Figure 4l). Produced as small bud in most specimens, as bilobed structure in one specimen.
Pereopod 2 (Figure 6b). Seven-segmented; coxa with two to six setae; basis with zero to two setae; ischium with four to six setae; merus with 16-18 setae; carpus with six to eight setae; propodus with 19-22 setae; dactylus with 15-17 setae and few small spines on inner margin; exopod reduced with zero to four setae, or absent.
Pereopod 3 (Figure 6c). Seven-segmented; coxa with zero to three setae; basis with zero to three setae; ischium with four to six setae; merus with 9-11 setae; carpus with five to six setae; propodus with 13-16 setae; dactylus with 12-14 setae and six spines on inner margin, and one small spine on outer margin; exopod absent.
Pereopod 5 (Figure 6e). Seven-segmented; coxa with zero or one setae; basis with zero to two setae; ischium with four to five setae; merus with five setae; carpus with three setae; propodus with 19-22 setae and two small unarticulated spines; dactylus with seven to nine setae and row of small unarticulated spines; exopod absent.

Discussion
Although there appear to be some differences between wild-caught larvae of Axianassa from the southwestern Gulf of Mexico (Ngoc-Ho 1981), larvae of A. australis described from Brazil (Rodriguez and Shimizu 1992) and the presently reported larvae of A. australis from the northern Gulf of Mexico and Atlantic coast of Florida, numerous similarities (Table I) suggest that these larvae represent the same or very closely related species. This is supported by preliminary comparisons of gene sequences of adult populations from these four areas (Felder 2001;R. Robles and D. Felder, personal communication). It was previously reported that ZI larvae of A. australis from Brazil had a single orange chromatophore on the mid-anterior section of the carapace (Rodriguez and Shimizu 1992). This would differ strikingly from the chromatophore patterns we observed in Texas and Florida populations of A. australis, which include large, red, paired chromatophores on the telson and fifth abdominal somite, as well as smaller, paired chromatophores on the lateralproximal ends of the carapace (Figure 2). However, recently obtained photographs of larvae hatched from a Brazilian female specimen identified as A. australis (F. Mantelatto, personal communication) indicate that these zoeal stages also have the distinctive paired chromatophores characteristic of the larvae we reared from the northern Gulf of Mexico and Florida. There thus appears to be little or no difference in carapacial chromatophore patterns.
The long dispersal phase of A. australis (likely to be at least 1 month) could certainly account in part for its extended distribution in the western Atlantic. The larvae of A. australis appear to be very close to those reported by Gurney (1938) as ''Laomediid D.I.'', which might be interpreted to suggest an even broader distribution. Among materials that Gurney assigned to this form were specimens from Samoa, the Great Barrier Reef and the western Atlantic, the latter from off the north-east coast of Brazil and apparently the materials he illustrated. Rodriguez and Shimizu (1992) noted that the ZI they described for Table II. Important similarities and differences of Axianassa australis with members of the family Ubogebiidae.
Differences from upogebiids (Ngoc-Ho 1981)  A. australis was similar to Gurney's description except for theirs lacking a spine on the antennal protopod and rudimentary pereopod 1 that occur in the ZI of ''Laomediid D.I.'' Both of these characters are present in the ZI of A. australis from Texas and Florida, as well as in larvae of Axianassa sp. described by Ngoc-Ho (1981), although a rudimentary pereopod 2 was also noted in the latter description. However, neither our specimens nor those of Ngoc-Ho have pleural spines on somites 2-4 at any zoeal stage, and these are conspicuously evident in the illustrated materials of ''Laomediid D.I.'' It is likely that these pleural spines are present in the ZI of ''Laomediid D.I.'', as no mention was made to the contrary, but they do not appear in the ZI of A. australis illustrated by Rodriguez and Shimizu (1992). Gurney (1938) specifically referred to the pleural spines on somites 2-4 of ''Laomediid D.I.'' as one of few features distinguishing it from otherwise very similar materials reported by Menon (1933) from Madras. Thus, it is doubtful that Gurney's materials are conspecific Table III. Comparison of first zoeal stage in Axianassa australis, Laomedia astacina (Fukuda 1982) and Naushonia crangonoides (Goy and Provenzano 1978), and Jaxea novaezealandiae (Gurney 1924;Wear and Yaldwyn 1966).
with A. australis, and there is no reason to conclude from his reports that the latter species ranges into Pacific Ocean waters. While A. australis, in the absence of pleural spines on somites 2-4 and several other features, does resemble the larvae from Madras that Menon (1933) tentatively placed with the upogebiids, it differs in a number of other characters (summarized by Ngoc-Ho 1981, p 251) from those Indian Ocean materials. Nonetheless, the occurrence of such similar larvae in Pacific and Indian Ocean waters suggests a broad distribution for close relatives of A. australis, which has yet to be established on the basis of adult materials. The genus Axianassa has at times been placed in the monotypic family Axianassidae or alternatively grouped with several other genera in the Laomediidae (see Kensley and Heard 1990 for review). Ngoc-Ho (1981) placed her larval specimens of Axianassa sp. into a group separated from other genera of laomediids (Naushonia, Jaxea, and Laomedia) on the basis of their distinct morphology. Major defining characters include the presence of lateral spines on abdominal somite 5, an appendix interna on the pleopods of the decapodid (presumed by Ngoc-Ho, demonstrated in present study), and an apical spine on the antennal scale in Axianassa, all of which can serve to distinguish this genus and thus separate the Axianassidae from the Laomediidae s.s. The separation is also apparently supported by lack of transverse uropodal sutures in the postlarvae of Axianassa, a feature present in all known postlarvae of the true laomediids s.s.
It would appear that larval development of Axianassa has almost as many similarities with species of the family Upogebiidae (Table II) as it does with those of the family Laomediidae s.s. (Tables III, IV), which accounts for why Menon (1933) placed larvae of this form in the former group. Recent molecular genetic studies also argue for phylogenetic placement of Axianassa in an intermediate position between these families, and thus for resurrection of the family Axianassidae (Tudge and Cunningham 2002). As now known, larval morphology clearly supports this conclusion.

Carapace
With supra-antennal spines and identified the parental female of those specimens, while Rafael Robles shared some preliminary molecular findings. Support for field and lab activities was provided to D. L. Felder from the US Department of Energy (grant no. DE-FG02-97ER12220), the US Geological Survey (cooperative agreements 00CRAG0009 and 00CRAG0035), and supplemented by several small project grants from the Smithsonian Marine Station. This is contribution number 105 from the Laboratory for Crustacean Research at the University of Louisiana at Lafayette, and contribution number 609 from the Smithsonian Marine Station. Grant no. DEB-0315985 from the US National Science Foundation.