Four species of cave weta identi fi ed from Te Paki Ecological District, females on left, males on right. 

Contexts in source publication

Context 1

... individuals of the same morphogroup, including males and females, which allowed them to be matched. Presence/absence of six apical spines varied consistently among the four species: fore femur retrolateral apical spine, mid femur prolateral apical spine, fore tibia superior prolateral apical spine, mid tibia superior prolateral apical spine, hind tibia inferior subapical prolateral spine and hind tibia inferior subapical retrolateral spine ( Table 1, Fig. 2). Together, these provide a combination of character states that consistently and reliably diagnosed adults and near adults of the four taxa encountered. Data from 319 adult TPA, TPC and TPD, 104 large TPB pre-adults and 154 other unidentified small juveniles were used in metric analysis to compare shape and size between sexes and among the four taxa. Only for TPA did we find any significant size difference between the means of adult males and females ( t -test, P < 0.05, Table 2). Although males and females of the other putative species did not differ significantly in size we note that there were quantitative differences between sexes. For instance, the hind femur of male TPC specimens was longer than the females ’ (Fig. 3). Adult TPB may well differ but this could not be tested with the present sample. Tukey HSD comparison tests showed there were significant differences in size of all three metric variables measured between TPA and TPB ( P -value < 0.0001), and between TPB and TPD ( P -value < 0.0001). In all cases, TPB was the largest even though no adults of this taxon were present in the sample. However, size alone was not reliable for distinguishing all taxa and the use of spine combination characters and subgenital plate shape was essential (Table 3). Including all spine data and the three metric variables ( P , HF and HT) PCA resolved numerous clusters when unidentified juveniles were included (Fig. 4A). TPA formed a cluster with a relatively large range in male and female sizes, but no discrete clusters indicative of instars or juveniles versus adults. TPB size variation was ...

Context 2

... their relative abundance over space and time. All statistical analyses used R v3.0 (R Core Team 2013). We used generalised linear modelling (GLM) assuming a Poisson distribution of count data to compare cave weta abundance over time. Site and habitat are nested variables. Strictly, ‘ site ’ is a random effect but was not fitted in this way because we wanted to determine whether there was additional site – site variation, and so we added terms sequentially into the model. In general, the number of trap-days was constant for all sites in a given month and variation among months was accommodated by including month as a factor in the models. Zero counts were inserted where traps were in use but caught no cave weta, while traps not in use for a particular month were coded as NA. We modelled cave weta abundance with and without data for the unidentified juveniles as a fifth taxon. The response variable was the count for each species. The explanatory variables were site, habitat type (native forest, pine or shrubland) and time (month of the year trapped), and all possible interactions of these variables were considered. Using the larger cave weta that were either adults or near adults, we identified four morphotypes from their spination, overall size and appearance. Males and females could be reliably paired based on spine counts and overall size. Despite their size, it was apparent that the largest individuals, all of which were assignable to taxon TPB, included no adults. Females, for instance, had relatively soft, pale and blunt ovipositors, whereas adult Rhaphidophoridae ovipositors are sclerotised, dark and pointed, often with fine serrations. The size range of the TPB specimens was greater than the adults of TPA, suggesting that a number of instars were included among the TPB sample. TPC and TPD were paler than TPA and TPB, and each had a unique spine combination (Table 1). TPC was similar in size to TPA and was initially overlooked because its pallid colouration and delicate form made it look like a juvenile of another species. The fourth phenotype, TPD, was the smallest of all of the putative taxa and this species also had a pale, juvenile-like appearance. Careful examination of the ovipositors in the females of TPD specimens revealed otherwise. The unique spine combinations of all four taxa reinforced our hypothesis that they were different species and not just variants of the first morphotypes encountered (Fig. 3). Initial assessment of the reliability of classifying juveniles into one of the four morphotypes indicated that it was prudent to exclude small juveniles (< c. 10 mm) from subsequent morphological analysis. There was no clear discrimination of juveniles based on size into one of the four putative species (Fig. 4A). We also found the presence/absence of some spines to be more variable among the small juveniles than adults, suggesting an ontogenetic component to spine development. Juvenile females could usually be distinguished readily from their respective adults by their pale, soft and blunt ovipositors lacking serrations. Juvenile males were less readily distinguished as late instars were very similar to adults. However, juveniles tended to be smaller, paler, with fewer apical and lateral spines on the legs, and the genitalia were underdeveloped and often difficult to see. Distinguishing penultimate and last instar individuals was the most difficult and in many cases impractical, but also may not be necessary except when secondary sexual characters are expressed only in the final instar. In order for species diagnosis to be as robust as possible we removed all small juvenile and uncertain individuals from the sample sets for each morphotype, excepting those large but not adult individuals of TPB, which could always be reliably distinguished from the other morphotypes. Males and females within the same species had very different subgenital plates associated with their different genital structure and function. From examination of adult specimens (and near adult TPB individuals) collected in this study we were able to identify eight distinct subgenital plate shapes, four relating to females and four to ...