Styloid crystals, presumed to be calcium oxalate, seem to be a fundamental characteristic of Iridaceae based on previous reports of their presence in several genera of both major subfamilies, and supplemented here by further records in 75 of the ca. 85 genera of the family. Styloids have now been reported in some 300 species of Iridaceae, of about 320 examined. These elongated crystals are lacking in a few scattered species but perhaps significantly from Sisyrinchium and some closely allied genera. Families and isolated genera possibly related to Iridaceae, including Geosiris and Isophysis, were also examined for crystal types. Styloids are present in Isophysis. Tecophilaeaceae and Campynemaceae have raphides and three species of Colchicaceae and several of Uvulariaceae have crystal sand. A few of the latter have raphides, sometimes with crystal sand. Crystals are absent in Geosiris. The difference in crystal types taken together with some significant differences in morphology suggest that Campynemaceae and Colchicaceae are not immediately allied either to Iridaceae or to one another. We suggest that Campynemaceae may be better placed close to Melanthiaceae or Burmanniaceae. Isophysis, with three stamens but a superior ovary, is probably best treated in its own subfamily of Iridaceae. Tecophilaeaceae are no longer believed to be closely related to Iridaceae and their placement in a different order is supported. Data on the kinds of crystals of the calcium oxalate type (raphides, styloids, and crystal sand) occurring in plant tissues are widely scattered in the literature and information concerning their distribution in the plant kingdom is not readily available to systematists. Although little is known about the function of such crystals in plants, their shape and location are often very characteristic at different taxonomic levels (Franceschi & Horner, 1980). A brief mention by Metcalfe (1961) that styloids (pseudoraphides) were particularly characteristic of Iridaceae seemed intriguing and worth further investigation to establish, as far as seems reasonable, the frequency of styloids and possibly other crystal types in the family. We have also surveyed the crystal characters in putative relatives of Iridaceae. These include the monotypic Madagascan Geosiridaceae; Colchicaceae (Liliaceae-Colchicoideae); Uvulariaceae (Liliaceae-Uvularioideae); the Tasmanian Isophysis; and the poorly known Australasian Campynema and Campynemanthe. The latter two genera are variously assigned to Hypoxidaceae, Colchicaceae (Dahlgren & Clifford, 1982), or to a separate Campynemaceae (Dumortier, 1829: 57-58; Dahigren & Rasmussen, 1983) and have been suggested to be close to Iridaceae (Dahigren & Rasmussen, 1983: 369-372). Uvulariaceae sensu Dahigren & Rasmussen, postulated as ancestral to Colchicaceae and to Iridaceae (Dahlgren, pers. comm.), comprises Hexacyrtis and the Disporum group of Liliaceae and is largely North Temperate. A few members of Tecophilaeaceae were also examined because this family was proposed as close to the ancestral line of Iridaceae by Hutchinson (1973), although there is little current support for his view. MATERIALS AND METHODS Dry or FAA-fixed living leaf samples were gathered from at least one to a few species of many genera of Iridaceae from both subfamilies, Iridoideae (here including Sisyrinchioideae) and Ixioideae. Samples were examined at Missouri Botanical Garden (MO) and Royal Botanic Gardens, Kew (K). At MO samples were sometimes cleared in 5% NaOH for several hours or more until satisfactorily transparent or more often were cleared in household bleach. They were then ' Supported by grants DEB 78-10655 and 81-19292 from the United States National Science Foundation. I thank R. Keating for technical advice and for helpful comments; R. Dahlgren for reviewing the manuscript and for suggestions including the disposition of Campynemaceae; and John Dwyer for his independent examination of the flowers of Isophysis. 2B. A. Krukoff Curator of African Botany, Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. 3Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166. 4Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey, Great Britain. ANN. MissouRI BOT. GARD. 71: 1013-1020. 1984. This content downloaded from 157.55.39.35 on Wed, 31 Aug 2016 04:36:22 UTC All use subject to http://about.jstor.org/terms 1014 ANNALS OF THE MISSOURI BOTANICAL GARDEN [VOL. 71 mounted immediately in glycerin or dehydrated through an alcohol series and permanently mounted in Canada balsam. At K samples were sectioned using a Reichert sliding microtome. Sections were stained in safranin and alcian blue, dehydrated through an alcohol series and mounted in Euparal. The samples were viewed between polarizing filters to detect the presence and type of crystals in the tissue. The sole species of Isophysis and Geosiris, one each of Campynema and Campynemanthe and several of Colchicaceae, Uvulariaceae, Liliaceae (sensu Dahlgren & Clifford, 1982) and Tecophilaeaceae (Table 1) were treated in the same way at MO for comparison. The genera and species studied are listed in Table 1 together with the crystal types observed. Voucher or accession number information is available from the authors but is not reported