TAXONOM1C STATUS OF PSITTACULA INTERMEDIA 



39 



juv. H- 

 Eastern F HI* 30 

 Western F -"8^-20 

 Sane Q 2 

 Sedgemore -^K5 



BMNHDl 



C — ^ — 40 

 R— «^- — 35 



-2 ^1 1 2 

 Factor 1 92.1%, 15.8; bill size 



Fig. 3 Graph of means (squares), standard deviations (heavy bars), and 

 ranges (narrow bars) of individual scores for putative parental species 

 and hybrids on Factor 1 , the only significant axis in a PCA of culmen 

 length (component loading 0.95), maxilla height (0.96), and maxilla 

 length (0.97). The number to the right of the range bar is n. Symbols are 

 as for Figure 1 and 2; eastern and western finschii are included as 

 separate groups. 



larger than intermedia. In the plot of the distal width of rectrix 2 (R2) 

 vs. the pale tip length of that feather (Figure 2C), intermedia fall 

 belweencyanocephala/roseata andhimalayana, but not mostfinschii. 

 Psittacula krameri is unique in its combination of a broad, very short 

 pale tail tip, and in this Sane's specimen is intermediate between 

 krameri and cyanocephala/roseata, and certainly not himalayana. 



On a PCA of three bill measures (Figure 3, Table 3) selected to 

 allow inclusion of as many specimens and live hybrids as possible, 

 the only significant axis was Factor 1 , a very strong size axis. On this 

 axis, cyanocephala and roseata had the smallest mean factor scores, 

 with the BMNH intermedia slightly larger. The other putative 

 intermedia and known hybrids fell between these and the succes- 

 sively larger finschii and himalayana groups. Psittacula krameri 

 was much the largest, and Sane's specimens were the largest of the 

 putative hybrid groups, again showing the influence of krameri. 



In a PCA for which variables were selected to allow inclusion of 

 one of Sane's specimens (Figure 4A, Table 3), the AMNHintermedia 

 and Sedgemore's birds group near each other, and between the 

 widely spaced roseata/cyanocephala and himalayana groups, but 

 overlap considerably with finschii. Sane's bird, however, falls 

 between the roseata/cyanocephala and krameri groups. 



Another PCA for which the variables selected allowed inclusion 

 of Sane's other specimen (Figure 4B, Table 3) showed AMNH 

 intermedia grouping out halfway between cyanocephala and 

 himalayana, with the mean of roseata falling out more distantly. In 

 this case, the second Sane specimen falls out much closer to 

 cyanocephala than to krameri. 



Wing and tail formulae 



In mean shortfalls of each primary tip from the wingpoint (Figure 5 A), 

 the AMNH intermedia are completely intermediate between 

 himalayana and cyanocephala. However, BMNH 1980.3.1 is very 

 like the mean of cyanocephala in pattern of primary shortfalls from 

 the wingpoint. AMNH intermedia are closer in mean primary short- 

 falls to roseata than to cyanocephala (and thus less intermediate 

 between roseata and himalayana; Figure 5B), but neither roseata 

 nor finschii could be ruled out as parental species on this basis alone. 

 Sane's single specimen on which these characters are measurable is 

 nearly intermediate in primary shortfall pattern between 

 cyanocephala andkrameri (Figure 5C), although again these data do 

 not rule out some other parental combinations. 



On mean widths of primaries, AMNH intermedia were again 

 intermediate between himalayana andcyanocephala except in width 

 of P2, a measurement that is highly dependent on shortfall of P3 

 (Figure 6A). In spacing between tips of rectrix pairs 3-6 (Figure 6B), 



T3 



tb 



> 



Oh 



43 



6C 



o 



u 

 © 



M 



fa 



.— 



1 



2 





— 







>5 





r~ 





r^ 



-1 



~ H 





<S 





u 



-? 



O 





+-. 





w 





M 





fa 



-3 



-4 



Bill, wing 1, R2 



-2-10123 

 Factor 1 71.5%, 4.3; size (R2 w, tip 1 uncorrelated) 



-2-10 1 2 



Factor 1 12.1%, 63.7; size (P3s, Plw uncorrelated) 



Fig. 4 Identity of AMNH intermedia with Sedgemore's hybrids, 

 distinctness from Sane's specimens, and intermediacy of all the above 

 between putative parental species: graphs of individual scores on 

 Factors 1 and 2 from PC As on measurements of adult males of putative 

 parental species and hybrids. Symbols are as for preceding figures. 

 Summary statistics of PCAs are given in Table 3B and C. (A) Variables 

 chosen to allow inclusion of Sedgemore's hybrids and Sane's first 

 specimen; (B) variables chosen for inclusion of Sane's second specimen. 



