572 A NATURALIST IN THE PACIFIC 



But, apart from this, the genus Quercus finds in its own constitution or 

 habit the greatest obstacle in most species to the adoption of a littoral 

 station. However, there are exceptional tendencies displayed by the ever- 

 green oaks ; and this is very significant, since in their xerophilous leaves 

 they possess the preliminary qualification for a station near the sea. 

 Quercus ilex, it is well known, shows a partiality for the sea-air, and 

 Q. virens, the " live oak," flourishes near the sea in the southern states of 

 America, a maritime variety being distinguished by botanists. One of the 

 willow-oaks of America, Q. phellos, which grows in swampy land, also has 

 a beach variety. 



The Hazel-tree (Corylus avellana) must be placed in the same category 

 with Quercus. I found the empty nuts commonly amongst the stranded 

 drift of the Sicilian and English beaches. The fruits were also frequently 

 noticed by Dr. Sernander in the Scandinavian sea-drift ; but he says 

 nothing of their empty condition. Mr. Darwin remarks, in the Origin of 

 Species, that he found that fresh hazel-nuts sank, but that after drying 

 a long time they floated for ninety days and subsequently germinated. 

 The floating-power is no doubt due to the cavity arising from the shrinking 

 of the kernel, and it is to this cause that Dr. Sernander attributed the 

 slight initial buoyancy observed by him. However, the hazel, like the 

 common oak, lacks the habit that would fit it for a station by the sea, and, 

 whatever capacity its fruits may possess for dispersal by currents, it is quite 

 useless for the spread of the species. 



NOTE 49 (page 131) 



On the Distribution of Ipomea pes capr/e, Convolvulus 

 soldanella, and convolvulus sepium 



Whilst Ipomea pes caprse is cosmopolitan in the tropical zones, Con- 

 volvulus soldanella is cosmopolitan in both the north and south temperate 

 zones ; but, as might be expected, the two species at times meet and their 

 areas overlap. Thus, according to Mr. Cheeseman (Trans. New Zealand 

 Inst., xx., 1887), they meet in the Kermadec Islands, in the South Pacific, 

 in about latitude 30 . From my observations on the coast of Chile it 

 would seem that C. soldanella in its northward extension fails somewhere 

 between Valparaiso and Coquimbo, that is to say, between 33 and 30 

 S. lat. Gay merely refers to the plant as existing in North Chile, which in 

 his time would include the coast between ^ and 24 S. lat. It intrudes 

 within the " thirties " on the coast of California and is found in Madeira in 

 about 33 N. lat. Ipomea pes capras in its turn extends into subtropical 

 regions, being recorded from the Kermadecs, as above noted, and from the 

 Bermudas in 32 N. lat. Owing probably to special physical conditions of 

 the coast, which are referred to in Chapter XXXII., this plant is evidently 

 limited to the tropics on the west coast of South America. It did 



