Reworking Family Trees for Lories

Abstract

L ories and lorikeets (Loriinae:

Psittacidae) are brilliantly colored parrots that feed primarily on nectar, fruits, and pollen (Forshaw 1973, Low 1977, 1998) and probably insects (Low 1977). Their digestive tract possesses special adaptations for a life of nectivory (Richardson & Wooller 1990), including a remarkable brush-tipped tongue (Salvadori 1891). Lories range throughout islands and archipelagos of the southern Pacific Ocean, where they are challenging to study in the field due to difficult terrain and remote locations. However, lories are accessible because they are widely bred in captivity by zoos and private aviculturists throughout America and many parts of the world.

Even though lories are easily distinguished from other parrots by their brush-tipped tongue, distinct beak shape, and feeding habits, they are very similar to each other and may have recently evolved into distinct species (Christiclis et al. 1991). Initially, lories were classified into genera (plural for "genus") primarily based upon external features and plumage color patterns (Salvadori 1891). Despite widespread criticism of the original family tree, this arrangement is still used by others (Peters 1937, Forshaw 1973). However, the lack of truly definitive characters has resulted in confusion; some lory species have been placed into 5 or 6 genera during the past 150 years and, surprisingly, into two or three genera since Peters' "checklist" 0937) was published. Thus, 101y classification remains a topic of lively debate among ornithologists and aviculturists. This lack of knowledge presents many problems for both conservation and ornithological pursuits. In this article, I will present

a brief overview of some of the taxonomic difficulties that exist for lories and propose a solution: construction of a molecular (DNAbased) phylogeny of the Loriinae.

Currently, most taxonomists divide lories into 8, 9, 11 or 12 genera containing 53 or 55 species. Recently, a combination of traditional and new methodologies indicates these classification schemes should be revised. For example, according to Joshua (1994),

whose recent taxonomic study of lories was based upon analysis of chromosome structure and number (karyotype data), there are three major Jory lineages. One Lineage is comprised of the two genera Cbarmosyna and Vini, while the second group includes Glossopsitta, Tricboglossus, Cbalcopsitta, Eos, Lorius, and Neopsittacus and the third distinct karyotypic form is Pseudeos, a rnonorypic (single species) genus.

Morphological and behavioral similarities suggest that Charm.osyna and vzni could be combined into one genus (Amadon 1942). Additionally, it is interesting to note that Pbigys and Vini share skeletal structures that are unique from other psittacines, suggesting that Pbigys could be subsumed into Vin: based on these features (Steadman & Zarriello 1987). Further, karyorype data, which is lacking for Pbigys, reveals striking similarities between Vini and Cbarmosyna, implying early evolution of this lineage away from other lories (loshua 1994).

The second lineage comprises the majority of Jory genera and therefore, it also is the most confusing. Traditionally, Glossopslita was distinguished from Tricboglossus by their

black beaks, elongated first primary and smaller overall size and Psitteuteles was separated on the basis of their predominantly green plumage and different color patterns (Mivart 1898). However, many experts argue about the validity of Glossopsiua and currently Psitteuteles is not generally recognized as a separate genus. Further, biochemical work indicates that Glossopsitta and Psitteuteles should both be included in Trichoglossus (Christidis et al. 1991). Indeed, this taxonomic ambiguity extends throughout many of the other Tricboglossus as well. For example, Trichoglossus haematodus is comprised of so many distinct geographically isolated varieties that it is the most intricate superspecies complex among the psittacines.

To my knowledge, Cbalcopsitta, Eos, Loriu.s and Pseudeos have never been carefully studied, however, they are considered by many to be very close. This confusion also extends to me species level for these genera, particularly for Cbalcopsitta species. According to Diamond (1972), me three Cbalcopsitta native to New Guinea comprise a "superspecies ring" of poorly defined taxa ranging throughout me lowlands. The different forms apparently result from regional contact being broken and reestablished a number of times between these semispecies: C. duiuenbodei on the north coast, C. atra in the far west and c. scintillata from me south coast and Am island. In fact, it is likely that the unique type specimen, C. spectabilis, is a naturally occurring hybrid between C. atra insignis and C. scintillata scintillata (Low 1998). The fourth Cbalcopsitta species, the entirely reel C. cardinalis, is found on the Solomon Islands instead of New Guinea and is often included in Eos clue to similarities in karyotype (Joshua 1994) or plumage color and pattern (Forshaw 1973). Similarly, Eos and Lorius are also poorly defined.