1909] 



OLSSON-SEFFER— VEGETATION ON SEA SHORES 87 



tion period extends over the whole year, but the sand drifts here much 



summer 



when wet the coherence of the sand is greater. Plants like Ammo- 

 phila, Cakile, Salsola, and Atriplex, of which the parts above ground 

 die even- winter on the Baltic coasts, grow all the year round on the 

 shores of Australia, and there is no difference, external or internal, 

 in structure. Convolvulus soldanella on the coast of Holland does not 

 differ from the same species on the shores of tropical countries, where 



(fig- 



warm countries more 



because of the longer periods of drought, which favor the drifting of 

 the sand. We refer to the enormous areas of sand formations in 

 Africa, both on the coasts and in the interior, in the deserts of Asia, 

 on the coasts of India and Australia, in the interior of the latter con- 

 tinent, on numerous tropical coasts as Java, Hawaii, etc. The dunes 

 which occur in really cold countries, as in certain parts of the United 



former. 



comparison 



It would be of considerable interest to have correct data of tempera- 

 ture conditions from the various localities where the author has made 

 his studies of the coastal sands. The field-work on which this paper 

 ls based » however, has been conducted for comparatively short periods 

 at each place, and the temperature observations made 



comparisons 



Official data, obtained 



atisfactory 



our special purpose, because the instruments usually are kept close to 



from vegetation . in more 



—t,^, *n ay irom vegetation, in more or less snenc-icu ^iu U »-, 

 and because of these facts the observations cannot serveior any definite 



conclusions as to 



fetation has developed. It is 



temperature 



necessary 



kind referred to. The 



author has compared a great number of temperature statistics from 



lff erent coasts, but he has entirely failed to find any apparent rules 



applicable to the development in general of sand vegetation in differ- 



wh-t imateS ' This ne S ative result is not due to absence of such lawS ' 



defi Ceitainly must exist li mcrel >' shows that our knowledge is 

 e c »ent and the present method of taking temperatures at meteoro- 



