LIFE CYCLE OF PARACYCLOPS 



69 



the segments of the female antennule and of the antennules 

 throughout the copepodid phase of development. These results 

 are presented in schematic form in Figure 23 which visually 

 indicates homologous segmental boundaries expressed in 

 successive copepodid stages. The aesthetasc derived from 

 ancestral segment XXI provides a reference point throughout 

 the copepodid phase: from the third segment of the first 

 copepodid, to the fifth segment of the adult female and the 

 fourteenth segment of the adult male. This, in concert with other 

 setation features, confirms the unequivocal identification of the 

 XX to XXI articulation (marked by the large arrow) in all stages. 

 Identification of the homology of the more proximal segments 

 relies on the presence of 1 1 setation elements on the third 

 segment of the male copepodid V stage. These 1 1 elements 

 testify to the segments derivation from 6 unseparated ancestral 

 segments (XV-XX) and indicates that the boundary between the 

 second and third segments represents the XIV to XV 

 articulation. Similarly, the presence of 5 setae on the second 

 segment indicates that the boundary between first and second 

 segments represents the XI to XII articulation. The remaining 

 segmental homologies in all stages, including the adult female, 

 can now be identified by tracking vertically through the male 

 stages back to copepodids I and II and then forwards again to 

 the adult female. 



The seventh segment of the female antennule is identified as 

 homologous with a single ancestral segment (XXV). The 

 presence of 3 setae on this segment therefore requires 

 explanation since the basic armature of segment XXV is 1 

 posterior margin seta plus 1 seta and 1 aesthetasc on the anterior 

 margin (Huys & Boxshall, 1991). We suggest that one of the 

 anterior elements may be a setiform aesthetasc and this should 

 be investigated ultrastructurally. 



This is the first time that the homologies of all antennulary 

 segments have been identified in all copepodid stages of both 

 sexes of any cyclopoid copepod, with reference to the 

 hypothetical 28-segmented antennule of the ancestral copepod 

 (Huys & Boxshall, 1991). Within the family Cyclopidae 

 antennulary segmental numbers are frequently reduced and, 

 indeed, many genera are characterised by their low segment 

 numbers. The possible involvement of heterochrony in the 

 reduction of antennulary segmental numbers has already been 

 noted. Gurney (1933) suggested that 'the reduced numbers of 

 antennulary segments in some species may be accounted for as 

 the persistence of a larval character in the adult', pointing out 

 that during development the last copepodid (Co V) typically has 

 an 1 1 -segmented antennule and that this number is commonly 

 found in adults, as for example in Microcyclops. 



Several authors working on cyclopids were aware of the 

 importance of determining homologies between the different 

 generic segmentation patterns and made detailed comparisons. 

 Gurney (1933) reviewed earlier work by Claus (1893), Manfredi 

 (1923), Lucks (1929) and Gelmini (1928) on the sequence of 

 segmental subdivision during development in Cyclops species. He 

 presented a tabular system showing derivations of a 1 7-segmented 

 antennule and confirmed this pattern using his own data on C. 

 strenuus. The strict determination of ancestral homologies of 

 antennulary segments provides new insight into the rather 

 confused world of cyclopid systematics at the generic level. 



The scheme of segmental development (Fig. 23) indicates that 

 at the moult from copepodid II to copepodid III it is the second 

 segment that subdivides. This contrasts with the scheme 

 presented for P. fimbriatus by Gurney (1933) who showed the 

 first segment subdividing at this moult. We base our 

 interpretation on the positions and relative lengths of the setae 



on these segments. In particular a relatively long seta is 

 positioned at the anterodistal angle of the first segment in 

 copepodid II and 2 similarly long setae are present at the same 

 position in copepodid III. Since no other setae of this length are 

 present we interpret this as evidence of the constancy of the 

 boundary marked by this seta and therefore conclude that the 

 second segment has subdivided. 



The setation of the second endopodal segment of the antenna 

 increases progressively through the copepodid stages of P. 

 fimbriatus in a very regular manner. The first copepodid stage 

 possesses 4 setae distributed around the inner-distal angle of the 

 segment (Fig. 8A). The seta which is located on the angle is slightly 

 stouter than the other 3 and is here identified as seta VIII, using the 

 numbering scheme proposed for another member of the family 

 Cyclopidae by Boxshall & Evstigneeva ( 1 994). The more distal seta 

 is then identified as seta IX and the 2 more proximally located setae 

 on the inner margin, as setae VII and VI. These 4 setae are 

 presumably homologous with the proximal group of 4 setae 

 present on the margin of the unsegmented endopod of the sixth 

 nauplius stage (Fig. 3F). At each successive moult through the 

 copepodid phase one additional seta is added proximally on the 

 inner margin of the segment (Figs 8A-F). Thus at copepodid II a 

 fifth seta (seta V) is added, at copepodid III a sixth seta (seta IV), 

 and so on, until the final moult to adult (= copepodid VI) at which 

 the ninth and final seta (seta I) is added. 



The progressive development of the setation suggests that 

 reductions in numbers of setae on the second endopodal segment, 

 which are common within the family Cyclopidae, may be 

 interpreted as resulting from heterochronic events. Cryptocyclops 

 bicolor (Sars, 1863), for example, is a cyclopid with only 7 setae on 

 the second endopodal segment. This number is typical of the 

 copepodid IV stage in both Puracyclops and the presumed 

 ancestral cyclopid stock, and may be interpreted as evidence that a 

 neotenic event within the Cryptocyclops lineage has interrupted the 

 progressive addition of setae. 



The ontogeny of copepodid stages and examination the 

 patterns of leg formation offer some evidence for inferring 

 copepod phylogenetic relationships (Ferrari, 1988). A common 

 pattern of development for legs 1-4, exhibited by at least 20 

 genera, was recognized by Ferrari. The pattern of development 

 for the swimming legs of P. fimbriatus is in accordance with this 

 common pattern as follows: 



Legs 



1 



2 



3 



4 



N 



TB 



TB 







I 



1 + 1 



1 + 1; 



Tb 





II 



2+2 



2+2; 



1 + 1; 



Tb 



III 



2+2 



2+2; 



2+2; 



l + l; 



IV 



2+2 



2+2; 



2+2; 



2+2; 



V 



3+3 



3+3; 



3+3; 



3+3; 



VI 



3+3 



3+3; 



3+3; 



3+3; 



(Where N = nauplius; Roman numerals = copepodid stages; PB = 

 primary leg bud; 1 + 1 = reorganized leg with 1 -segmented exopod and 

 endopod; 2+2 = leg with 2-segmented exopod and endopod; 3+3 = leg 

 with 3-segmented exopod and endopod). 



Acknowledgements. We are grateful to Dr Rony Huys for his helpful 

 comments on drawing techniques and on the manuscript. S. Karaytug 

 would also like to thank Dr Steve Alston for his assistance in the 

 laboratory. This research has been supported by a postgraduate grant 

 from the University of Balikesir, Turkey to S. Karaytug. 



