Saffron crocus is widely, and probably correctly, regarded as sterile (at least in the sense that I have never seen reports of it setting viable seed). However, if the reason for the apparent sterility is that saffron crocus is triploid, then it might very well set viable seed if pollinated by a compatible tetraploid crocus. To this day the origin of saffron crocus is still uncertain. It's often said to have been derived from Crocus cartwrightianus. but there are other members of the sativus group which seem just as likely to me (C. moabiticus for example). One of the common ways for triploid plants to form is the failure of one of the parental gametes to undergo reduction division. This seems to happen under various conditions, but sometimes when the parental plants are of different but closely related species or breeding lines (thus the distinction made between autotriploids and allotriploids; in autotriploids the genetic material is from one breeding line, in allotrioploids the genetic material is from two lines)). What I've written above is old news - it was well known a half century ago. It's hard to believe that someone during that time has not tried to raise seed from Crocus sativus by inducing tetraploidy in a compatible species - if only to test the hypothesis or determine if Crocus sativus is autotriploid or allotriploid (or aeuploid). Or for that matter to duplicate the event which originally produced saffron crocus. Or maybe they have, and it's not widely publicized. Here's another aspect of this issue. If, as seems to be the case, Crocus sativus does not correspond to any true species (i.e. species in the sense of sexually reproducing population), then it is a "ghost species" - one which really does not exist. Since it cannot with certainty (or general agreement) be assigned to another species of the genus, a peculiar situation arises. Crocus sativus is the nomenclatural type for the genus Crocus. In other words, the genus Crocus has as its nomenclatural type a "species" which does not exist. The nomenclatural type of the genus Crocus is a clone, perhaps a hybrid clone, but a clone which is not a member of any true, sexually reproducing species. And as a clone not assignable to any species, it it should be classified at rank individuum, not at rank species. A similar situation exists with the genus Iris: its type is the Linnaean Iris germanica. But there is no such species as Iris germanica, except in the nothospecific sense. Iris germanica is not the name of a species but rather a name for a group of old hybrids. Jim McKenneyjimmckenney@jimmckenney.com Montgomery County, Maryland, USA, 39.03871º North, 77.09829º West, USDA zone 7 My Virtual Maryland Garden http://www.jimmckenney.com/ <http://www.jimmckenney.com/> BLOG! http://mcwort.blogspot.com/ Webmaster Potomac Valley Chapter, NARGS Editor PVC Bulletin http://www.pvcnargs.org /<http://www.pvcnargs.org/> Webmaster Potomac Lily Society http://www.potomaclilysociety.org/ <http://www.potomaclilysociety.org/>