OF THE SEUGIPE-ALAOoAS B^VSIN OF UIIAZIL. 417 



shedding the great precipitation of the region and sending it in powerful torrents down 

 the soft margins of the formation whence enormous quantities of earth are swept 

 away. 



(II) The sterility of the soils derived fi-om the tertiary is notorious. This steril- 

 ity is due largely to the tendency of the surface clays to turn the rainfall instead of 

 allowing the water to soak into the ground. But little water is able to penetrate the 

 ground, and for this reason dry weather is invariably fatal to crops in tertiary soils. 

 The natural vegetation is remarkably scant, except in the deep ravines. The wiiv 

 grass, stunted trees, and cacti that grow over these sterile plateaux, or "chapadas" 

 as they are called, offer but little resistance to the water flowing from the surface. 



(III) In the description of sections in the tertiary it may be seen that there are 

 no beds in the whole column capable of long resisting the ordinary processes of sub- 

 aerial erosion in the tropics. The beds are usually clays and sands variously commin- 

 gled. I have seen but one indurated rock in the series — a xcvy hard, glassy quail zitc, 

 metamorphosed in weathering, but this rock does not appear to form a continuous bed 

 at any one place, or to occupy a definite position in the series. In such soft and easily 

 eroded material the heavy rainfalls of Brazil do the greatest possible amount of work. 



(IV) Predjiitation. — That })reeipitation in the tropics is not as evenly distrib- 

 uted through the year as in temperate regions is a well-known fact. 



In Brazil the year is divided as naturally into rainj' weather, "tempo de chnvn,^'' 

 and sunny weather, "teinjyo de sol,^'' as in temperate regions it is divided into winter 

 and summer; but it is important to note also that the rainy season is the hot season, 

 the " tempo de sol," referring to continuous sunshine. 



The rainy season does not set in at the same time in all parts of the country, 

 but may begin one, two or three months later or earlier in one place than in anolluM-. 

 This, however, does not all'ect the general results. One of the best records of rain- 

 fall in Brazil of which I have knowledge is that kept by the St. John del Key Gold 

 Mining Company at Morro Velho, in the province of Minas. This record covers the 

 l)eiiod from 1855 to the i)resent. A resume of these observations published in 1880 

 shows* the average annual lainfall up to that time to be 1G37 millimetres, and that 

 89 per cent of this, or 1457 millimetres, fell during the six months from October to 

 March inclusive, while from April until September only 180 millimetres fell. 



A series of observations kept at the city of Sao Paulo during the year 187.)t 



* See " Relatorio do W. Milnor Roberts sobro o Uio Sao Francisco," Itio ilc Janciri), ISSn. .""1 /.'. r/«^l ,1,- EiK/m. 

 karid for May, 1880. 



t The Uio Newt, February 5, 1880. 



A. r. S. — VOL. XVT. .'Ja. 



