SEVEN NEW WHITE-WINGED DOVES 



21 



Rio Pocrf and the Rio Anton, Code. This race is believed to be non- 

 migratory. 



The breeding habitat in the mangrove swamp woodland apparently 

 is characteristic of this race. None of these birds was observed in 

 other habitat types in the localities visited. 



SPECIMENS EXAMINED 



Panama: La Isleta and Rio Pocri in Aquadulce, Province of 

 Code; and Rio Parita in Monagrillo, Province of Herrera. 



DISCUSSION 



Of the 12 subspecies mentioned, their grouping in the following 

 table, map, and figures is based on their breeding distribution in 

 three zones, (1) the Gulf and Caribbean, (2) the interior high- 

 lands, and (3) the Pacific Coast. Within these groups the arrangement 

 is from north to south. Those of group 1 breed in lowlands of the 

 eastern coast of Mexico and Central America and in the West Indies, 

 and include asiatica, peninsulae , and australis. They are of medium 

 to small sizes for this species, have shorter wings and tail, and live 

 chiefly in tropical and subtropical lowlands. 



Group 2 of the interior highlands includes grandis, monticola, and 

 alticola. They are large for whitewings, have longer wings and tail, 

 and do not occur in lowland localities. 



Group 3, Pacific Coast, includes mearnsi, palnstris, insularis, collina, 

 panamensis, and meloda. The races in this last group are of medium 

 to large size, and the ranges of several extend from the lowlands in- 

 land to higher elevations. Their wings and tail are of medium length 

 to long except in the mangrove-dwelling race panamensis, in which 

 these dimensions are smaller. 



Figure 1 shows the tentative boundaries of the breeding ranges of 

 the various subspecies, but further information will undoubtedly 

 result in many changes in this map. In the case of australis, for ex- 

 ample, the breeding range is without doubt more extensive than 

 shown, but in the series of birds examined, only the localities in north- 

 em Honduras were represented by breeding specimens. Their winter- 

 ing ranges are not mapped because the overlapping of asiatica and 

 mearnsi on the ranges of resident races is so complex. In some places 

 three different races may be wintering in the same locality. 



There are many unanswered questions in white-winged dove tax- 

 onomy and distribution, especially in areas where insufficient field work 

 and collecting have been done. To determine the ecological distribu- 

 tion of distinct populations, additional specimens should be taken 

 during the breeding season from Mexico south to Costa Rica. Likewise, 

 there are considerable gaps in our knowledge of the distribution and 

 taxonomy of this species in South America and the West Indies. 



