MONTIA FONTANA L. 



By S. M. Walters 



Botany School, Cambridge University. 



The Linnean species Montia fontana has long been recognised as very variable, 

 and many attempts have been made to subdivide the aggregate into smaller taxonomic 

 units. The earliest attempts were rather naturally based on the more obvious type of 

 variation in habit which was closely correlated with differences of habitat ; thus Gmelin 

 (1805) recognised two species, M. minor and M. rivularis, the former a small terrestrial 

 plant with erect stems, the latter a more or less submerged or floating aquatic plant 

 with long trailing stems. The less striking variation shown by the ripe seed was first 

 recorded by Chamisso (1831), whose Montia lamprosperma of N. Temperate and Arctic 

 Eurasia was distinguished by the possession of a smooth, shining seed, in contrast to 

 the dull, tuberculate seed type common in Central and Southern Europe. In most later 

 work, the seed-coat differences have been used in attempts to define the species within 

 the aggregate, usually in association with the habit characters. Thus Ascherson & 

 Graebner (1919) give three species in the aggregate : M. minor, annual, with dull, tuber- 

 culate seed; M. lamprosperma, annual, but with looser growth, and shining seeds with 

 very brittle coat; and M. rivularis, perennial, with shining, finely tuberculate seeds. 

 Koch's Synopsis (1892) gives the same three species, but with slightly different diagnoses. 

 The French and other W. European floras have generally adopted M. minor, the erect 

 land form with dull tuberculate seeds, and M. rivularis, more or less aquatic, with more 

 shining, finely punctate or tuberculate seeds. Lindberg (1901), in a general review 

 of the problem, gave as his opinion that any attempt to use characters of habit or vegetative 

 structure to define the taxonomic units within the aggregate was useless, and that the 

 characters of the ripe seed provided the only satisfactory basis. He agreed fundamentally 

 with Chamisso' s division into two main types, for which he adopted the names M. 

 fontana subsp. lamprosperma and subsp. minor; the M. rivularis of Continental authors he 

 considered to be merely an aquatic form of M. minor and described, as var. horeo-rivularis, 

 analogous aquatic forms of subsp. lamprosperma. Samuelsson (1922) considered the Scan- 

 dinavian types within the aggregate, and gave a good review of previous treatments by 

 Scandinavian workers; he agreed with Lindberg in distinguishing two forms on seed 

 characters, and in rejecting any characters other than those of the ripe seed, though 

 differing from him in his interpretation of M. rivularis. 



In Britain, the division into two types - the northern shining-seeded and the southern 

 dull, tuberculate-seeded types - had been generally accepted, until Beeby (1909) pointed 

 out that British material was more satisfactorily divisible into three, not two, seed-types. 

 These he called : subsp. lamprosperma (the northern type), subsp. minor var. chondrosperma 

 Fenzl (the southern type), and his new subsp. minor var. intermedia. Each of these could 

 exist in ' land and water states ' which were ' merely temporary conditions directly 

 induced by the environment, and not varieties.' Druce (1920) gave some details of the 

 British distribution of var. intermedia, and stated that he had seen material referable to 

 this from Belgium and Spain, and that it was ' doubtless elsewhere.' The existence 

 of this third seed-type, intermediate between the other two, has undoubtedly contributed 

 to the difficulties of British botanists attempting to apply continental work to the British 

 material, and the distinctness of the three types has not generally been realised (cf. however 

 Salmon (1931), who was obviously familiar with three types in the field). 



X 



