The chiefly North American Hypericum canadense has two isolated centres of occurrence in Ireland and one in the Netherlands. In the latter it was collected for the first time in 1909, subsequently in 1918 and 1932, but it was not recognized until 1934. Records are confined to northern Twente (eastern part of Overijssel, fig. 1). About 1935 it appeared to be locally rather common, but from 1940 onwards a strong decline followed and nowadays only one locality is left where H. canadense still occurs with certainty. Ever since its discovery in Europe the question has been asked whether this species would have occurred there of old or be a recent acquisition and in the latter case whether it was introduced by man. Those who have studied it in the field generally tend to consider it indigenous but hesitate to call it a relic. No indications of human interference in its arrival in Europe are known, the (former) remoteness of the centres of occurrence rather pointing to the contrary. Nevertheless a rather recent arrival is suggested by its apparent expansion in the first half of the 20th century, although this may have been connected with reclamations. Synecological data are also suggestive of a neophyte. Syntaxa in which H. canadense was recorded in Twente range from the Panico-Illecebretum (on coarse sand only wet in winter) to the Eleocharitetum multicaulis (on peaty sand just drying out in summer), the extremes having no other species in common (tabel 1). There are some notable differences with Irish relevé's containing a.o. much Carex echinata and C. tumidicarpa (Parvocaricetea). No indications of a strikingly narrow ecological amplitude were noticed, nor is H. canadense (hibernating by means of buds) as vulnerable and ephemerous in its occurrence as several therophytes of wet heathland, though it clearly has its minimum demands as to water supply. Several species are more or less analogous to H. canadense in having northern America as their main area and having been discovered in Europe only in the last few centuries. Apart from H. canadense the section Bathrys of Hypericum comprises four such species, and also in Juncus several examples are known like J. canadensis and J. tenuis. The latter two as well as Oxycoccus macrocarpos have filled up a hitherto ‘empty’ niche in wet heathland which must have been present for a long time. J. canadensis is especially similar to H. canadense in its habitat and likewise shows no indication of human introduction. Presumably H. canadense has arrived in the Netherlands rather recently and at first spread rapidly when reclamations created a suitable habitat on a large scale. Subsequently drainage, asphalting of sandy tracks and disappearance of nearly all moist heathland caused its disappearance but from one single locality.