Czech J. Genet. Plant Breed., 41, 2005 (Special Issue) Twenty-one Years Later: The Impact of Löve and Dewey’s Genomic Classification Proposal 1 2 M. E. B�������� and R. ��� B������ 1Intermountain Herbarium, Department of Biology, Utah State University, Logan, Utah, U.S.A. 84322-5305; 2Department of Crop Science, Swedish University of Agricultural Sciences, SE-230 53 Alnarp, Sweden, e-mail: [email protected] Abstract: The Fifth International Triticeae Symposium was held 21 years after Drs. A���� L��� and D������ R. D���� independently proposed that generic classification in the Triticeae be based on genomic constitution, spe- cies containing the same genomes being placed in the same genus. This was a startling proposal, particularly for taxonomists in North America where the existing treatments of the tribe still reflected the proposals made a century earlier by Bentham and Hackel. Although few taxonomists endorsed the principle of genomic circumscription, current generic treatments of the Triticeae have been strongly influenced by the wealth of genomic data available. The challenge that faces taxonomists is to determine which, if any, morphological characters or character sets can be used to identify the genomic composition of species in the tribe. The challenges for those engaged in molecu- lar research include elucidating the mechanism that underlies the formation of genomes and understanding the causal relationship of genomic constitution and morphology. Keywords: Triticeae; taxonomy; generic delimitation; genomic classification In 1984, Löve and Dewey independently proposed constitution of a species from its customary taxo- that generic classification in the Triticeae should be nomic treatment. One may not agree with Löve’s based on genomic constitution (L��� 1984; D���� taxonomic treatment (i.e. his decisions as to which 1984), in other words, that species with the same names represent good taxa and which should be genomic constitution be placed in the same genus. synonyms), but his synopsis is a phenomenal sum- The Fifth International Triticeae Symposium was mary of nomenclatural information for the Triticeae held 21 years after publication of their papers. This and is the obvious starting point for an electronic makes the papers “adults” according to the law in nomenclatural catalogue of the tribe, one of the some countries. It seemed appropriate to start this necessities for development of information on the symposium with a review of the impact of their Triticeae in the Global Biodiversity Information recommendation on research in the tribe. Facility (GBIF; http://www.gbif.org). The frequent- The two papers differed in their focus and taxo- ly heard statement that the tribe includes about nomic scope. Löve provided a checklist of the 350 specific or subspecific taxa is based, in large whole tribe that included all the names of which part, on Löve’s treatment. he was aware, together with their place of publi- Löve also provided short descriptions of the cation. As he stated in the introduction, many of genera and subgenera, including their chromosome Löve’s decisions were based on existing treatments number and genomic constitution. The descriptions rather than personal knowledge of the plants in- were not intended to be comprehensive, nor to serve volved and in some cases he inferred the genomic as the basis for writing keys. K������ (1989) used Proc. 5th International Triticeae Symposium, Prague, June 6–10, 2005 3 Czech J. Genet. Plant Breed., 41, 2005 (Special Issue) them as the starting point for a cladistic analysis Both Löve and Dewey stated that future work of the tribe, but added information from other might lead to the necessity of recognizing additional sources in order to develop an almost complete genera. In 1986, L��� published two additional data matrix. A more serious problem arises from genera: Trichopyrum for species having a genomic Löve’s failure to cite sources for the cytological constitution of EeSt or EeEeSt and Psammopyrum information included. Such an omission is quite for species with a LEe, LLEe, or LLEeEe constitu- common in floras and checklists. Unfortunately, tion (using designations of Wang et al. (1994) that in at least one case, Löve’s omission conceals a are equivalent to the JS, JJS, GJ, GGJ, and GGJJ, lack of supporting evidence: his statement that respectively, used by Löve). Dewey, although aware Leymus has an NsJ genomic constitution (genomic that Elymus, as he and Löve treated it, contained symbols used in accordance with recommenda- species with at least three different genomic con- tions of the International Triticeae Consortium, stitutions, was not prepared to publish additional see W��� et al. (1994); http://herbarium.usu.edu/ generic names to reflect the different genomic Triticeae/genmsymb.htm) was based on instinct groups. Since he did not accept Lophopyrum as dis- rather than cytological study (L���, pers. comm. tinct from Thinopyrum, it is unlikely that he would to D����, circa 1986). His statement was accepted have accepted Trichopyrum and Psammopyrum. by several individuals, including B�������� et Although Löve and Dewey worked completely al. (1983) and D���� (1984). Subsequent work independently, Löve was well aware of Dewey’s (Z���� & D����� 1991; W��� & J����� 1994; work and used it in formulating his ideas. Dewey A���������-J������ & B������������ 2001; became aware of Löve’s interest after circulating Ø������ & H�����-H������� 1994a, b) provides the manuscript of a paper he had presented at a no evidence for the presence of the J genome; the conference held in 1979 (see D���� 1982) in which second genome appears to be a modification of the he proposed a more timid version of his later clas- Ns genome. A well-documented online database sification. Löve wrote to Dewey, urging him to of cytological information needs to be developed be more courageous and telling Dewey about his so that, in future, it will be easy to determine for work on the Conspectus. Thus, although the two which species the genomic constitution has been worked independently of each other, they were determined and by what means. aware of each other’s work when they prepared D���� (1984) in paper, The genomic system of clas- their manuscripts. sificationas a guide to intergeneric hybridization with The immediate response of taxonomists to the the perennial Triticeae, was restricted to the peren- proposal made by Löve and Dewey that generic nial genera, among which he included Hordeum. circumscription in the Triticeae be based on ge- In it he stated that, because genomic constitution nomic constitution was negative. To understand was the best guide to biological and phylogenetic the response, it is useful to consider the generic relationships within the Triticeae, it should be used treatments in use in 1984. For this purpose, we for generic delimitation. He discussed the relation- review the treatment of the tribe in the major floras ships of the genera he recognized based on wide that were in existence at that time. hybridizations and classical analyses of meiosis and cited the papers documenting his statements. Perceptions of the Triticeae in 1984 He listed the species he knew to be included in each genus, but did not attempt to include species Americas: In 1984, the majority of North Ameri- with which he was not familiar. He did not provide can floras based their treatment of the Triticeae morphological descriptions for the genera, nor a (which many still called Hordeae) on the treatment complete list of synonyms. in H�������� (1951) which was, in turn, based on The only difference between Dewey and Löve B������ (1882) or H����� (1887). In other words, in their treatment of the perennial Triticeae was Agropyron included annual and perennial species Löve’s recognition of Lophopyrum, a genus whose with a tough rachis and a single spikelet at each members Dewey included in Thinopyrum. This node, species with tough rachises and more than difference reflected a difference in opinion as to one spikelet at a node being placed in Elymus. The whether the genome of T. elongatum [≡ L. elongatum] one departure from Bentham’s treatment involved should be designated JE or E (W��� 1985; W��� & recognition of the North American genus Sitanion, H���� 1989; J����� 1990; W��� et al. 1994). which differs from other species of Elymus in hav- 4 Proc. 5th International Triticeae Symposium, Prague, June 6–10, 2005 Czech J. Genet. Plant Breed., 41, 2005 (Special Issue) ing disarticulating rachises. In 1956, S������� com- Immediate Response mented that, genetically speaking, the whole tribe could be treated as a single genus. Despite this, The immediate response of most systematists American floras continued to adopt the traditional was negative. Almost all objected to the concept of generic treatment of the tribe (e.g. R��������� et basing generic delimitation on what was seen as a al. 1970; S������ 1978; B����� 1983). single character, pairing behavior of chromosomes. Eurasia: The former Soviet Union followed This objection reflects a long history of discovering T������ (1976), whose treatment differed from that new characters or approaches, despite their that of B������ (1882) and H����� (1887) in re- initial promise, have never lived up to their early cognizing several additional genera, both annual promise for solving taxonomic problems; exceptions and perennial, e.g. Eremopyrum, Taeniatherum, to a generalization are soon discovered. Why should Elytrigia, and Leymus. Agropyron was restricted to looking at meiotic pairing be different? Löve and A. cristatum and its allies (the crested wheat- Dewey argued that pairing behavior reflected the grasses), the other species Bentham included in genetic similarity along the chromosomes and was, Agropyron being placed in Elymus or Elytrigia. therefore, a summary of multiple characters. They The treatment in Flora Europea (M������� 1980) knew that, in some instances, pairing is overridden was similar to that in T������ (1976) except that by genes that control pairing but considered such Elytrigia was included in Elymus.