132 Washington — Submarine Eruptions of 1831 and 1891- 



ing in this connection. These areas of very shoal water are 

 all small, and usually set in groups on a rather shallow "bank." 

 In some cases, as the Keith reef, the bottom is so near the sur- 

 face of the water that the sea breaks. Although many of 

 these are covered with coral, the characters of their submarine 

 topography (shown in detail for some on Chart No. 2127), 

 which resemble those of shoals known to have been the sites 

 of submarine eruptions, lead to the conclusion that these shal- 



FlG. l 



low patches have also originated in the same way, and that 

 submarine eruptions have been of comparative frequency in 

 this portion of the Mediterranean. That we have direct 

 knowledge of so few of them need not cause surprise, when it 

 is remembered that many of them probably gave rise to 

 no solid or lasting island, as was true of Foerstner Volcano, 

 and that their periods of activity were often short, so that, as 

 was the case with the eruptions of 1845 and 1846, it is only by 

 the rare chance of some passing vessel (generally the small 



