I IO 



NATURE 



[April 3, 19 13 



sea are not always proofs of subsidence. Such valleys, 

 like coral reefs, may owe their existence to different 

 factors in different cases. 



The harbours of this coast are the high parts of 

 submarine fault valle) s, and the portion bordered by 

 land is often only .1 small part of the whole. Yet the 

 evidence is ver) clear that the coast has risen regu- 

 larly since the Red Sea was made, and that river 

 erosion has had nothing to do with the formation of 

 these steep-sided rifts. The coast of equatorial East 

 Africa is essentiall) similar, though there the fault 

 features are not so obvious, as' they have been 

 smoothed down somewhat by the heavy rainfall. Yet 

 such harbours as Mombasa^ Kilindini, Tanga, and 

 YVasin are obviously homologous with those of Port 

 Sudan. Suakin, and the desert harbours of the Red 

 Sea. 



Chwaka Bay, on the east coast of Zanzibar, is part 

 of a depression which runs right across the island in 

 a south-west direction, and forms long inlets where it 

 reaches the sea on the other side. 



The island of Pemba, a little north of Zanzibar, is 

 dissected in a wonderful way on its west side by long 

 fiords running far into the land, and ending in man- 

 grove creeks, perfect slaver- and pirate-hiding places. 

 The land is comparatively high and of great fertility, 

 so that a coasting voyage among the fiords is one of 

 great pleasure and interest. 



Surely these are typical drowned valleys? Not at 

 all; the island, like Zanzibar, was elevated in late 

 Tertiary times, the fossils of its coral cliffs are Pleis- 

 tocene or recent, and both islands have remained at 

 or near their present level while their reefs were 

 carved out. In places on the east coast of Zanzibar 

 the reef is up to three miles wide, and is something 

 between a fringing and a barrier formation. Yet it 

 was all formed by abrasion of elevated coral land. 



Along the edge of the deep water, in a line at right 

 angles to the Pemba fiords, is a regular but broken 

 line of reefs and islands, a typical barrier. Where 

 the islands have not been worn away bv the sea the) 

 are formed entirely of elevated coral, 'and the reef's 

 have been formed by abrasion of land. When this 

 process is complete, there will be a typical barrier 

 reef, should the islands continue stationary at the level 

 to which they were originally elevated. 



The two islands, Zanzibar and Pemba, were origin- 

 all) regular oval cups of coral (or perhaps saucers 

 rather, as in Pemba coral limestone is seen at lowest 

 tide level a long way up Chaki Chaki Bay), and were 

 connected with the mainland. These were filled with 

 the -and and pebbles which now form most of the 

 hills. After elevation they were cut off from the 

 mainland bv marine erosion, possibly (in the case of 

 Pemba at least) also by a comparativelv deep fault. 

 Other faults on this side cracked the 'saucers, and 

 irregularity was further induced bv marine abrasion, 

 in some places made excessive by powerful silt-bearing 

 currents, in others neutralised bv the protection 

 (ford d by growing coral and alga-. As along thi 

 deep water conditions more uniformlv favour the 

 growth of protecting organisms, the edges of the 

 saucers retain their regularity, while within the mass 

 ot sand and the limestone on which it lies is subject 

 to rapid degradation. The fiords are extending into 

 the land among the mangrove swamps at their heads. 

 The roots of the mangroves penetrate the crevices of 

 the coral and cause shallow accumulations of 

 mud, both factors for the disintegration and solution 

 of coral limestone. I take the extremely complicated 

 outline of the south end of Pemba Island to bi an 

 example of the barachois described by Prof. Gardiner 

 in the results of the Percy Sladen Trusl expedition to 

 the western Indian Ocean. 



NO. 2266. VOL. 91] 



In the Red Sea the heads of the harbours are being 

 filled in with blown sand and flood-borne alluvium" 

 but in Pemba there are no delta-like formations ; the 

 marine currents are far too powerful to allow of them, 

 even if the streams were big enough to convey any 

 quantity. of alluvium. 



In the Red Sea and equatorial Last Africa we thus 

 have, on stationary or rising coasts: — (1) Submarine 

 valleys which are scarcely altered from the original 

 fault rifts ; (2) those of similar origin somewhat dis- 

 guised by the effects of tidal currents and fresh-water 

 streams. 



The above have little if any continuation into the 

 land, but in Pemba are found : — (j) Long fiords simu- 

 lating drowned valleys, but of the same origin as (1) 

 and (2); (4) barrier reefs formed by abrasion alone, in 

 Zanzibar_and Pemba. 



In the Red Sea are (5) barriers which are features 

 of the sides of a rift valley merely coated with coral. 



(6) The barriers of the equatorial coast may be of 

 the same origin as those of the Red Sea, but, if 

 Zanzibar Island be reckoned a part of this barrier, 

 removal of the intervening land has been at least the 

 cause of the final separation, as proved by the fauna 

 of the island. Probably they are comparable to the 

 barrier of the west side of Pemba in structure and 

 origin, due to the faulting and erosion of the hetero- 

 geneous material of coral deposits. 



My statements here are dogmatic for want of space, 

 but details are given in my papers in Proc. Camb. 

 Phil. Soe., 1902, and Journ. Linn. Soc, vol. xxxi., 

 1907 and 1911. 



My being in a distant and isolated village, which is, 

 however, a few yards from coral reefs, is my excuse 

 for the delay in your receipt of this contribution to the 

 discussion. Cyril Crossland. 



Dongonab, Red Sea, February 24. 



Elliptical Lunar Halos. 



On the morning of January 26, 1908, while observ- 

 ing with the Keeler reflector of this observatory, I 

 noticed a curious lunar halo, which I described as 

 follows in the notes made at the time: — 



"At 4.25 a.m. (Eastern standard time) I saw an 

 elliptical halo close to and concentric with the moon. 

 Its major axis was vertical, and was about 7 long, 

 the lunar diameter being used to make this estimate. 

 The horizontal axis was a little less than half the 

 vertical, or about 3 . The halo lasted only about one 

 minute, but was unmistakable, being well defined 

 (except near the bottom) on both its inner and outer 

 edges. It reappeared less distinctly at 4.31 a.m.. 

 lasting only a few seconds. I watched the moon 

 until 4.50 a.m., but saw no third appearance. The 

 halo was white in colour. There was a light east 

 wind blowing at the time, and there was a little haze 

 in the sky. The temperature was 2S F., and had 

 been very constant all night. The moon was at last 

 quarter." 



At my request the other observers here have watched 

 for such a halo, and on December i, 1908, both Prof. 

 Jordan and Dr. Baker witnessed a much longer 

 apnearance of it. Prof. Jordan's notes read as fol- 

 lows : — 



"An elliptical lunar halo was visible here on thi 

 evening of December 1, 1908. The night began with 

 a very clear sky. but about eight o'clock a slight haze 

 became visible in the neighbourhood of the moon. 

 This soon began to take definite form, and the halo 

 developed. It remained visible until about 9 p.m., when 

 it gradually disappeared and the sky soon became cloudy. 

 The halo varied greatly in brightness because of the 

 varying amount of haze in the atmosphere : verv 



; 



h 



