The Coral Reefs of the Spermonde Archipelago (S. Celebes). 1. — The coral reefs of the Spermonde archipelago, West of Makassar (S. Celebes) are distributed over a submarine plateau, the so called Spermonde shelf, which to the landward side, gradually passes into the coastal plain of Makassar and Maros. To the East of Maros tertiary limestone mountains form the boundary of this plain; up till now we know Eocene and Miocene marine strata from these mountains. At the foot of these limestone mountains we find fossil coves about 30 meters above the present level of the sea, probably caused by a pleistocene abrasion and thus showing a pleistocene or postpleistocene negative shift of the strandline of about 30 meters. The coastal plain consists of volcanic material that was deposited in the sea, as is shown by foraminifera found in these volcanic tuffs (see fig. 1—3). The submarine plateau forms, together with the coastal plain, one gradationplane; we can reconstruct the following history of this region (see fig. 4): During the pleistocene a sinking of the sea-level took place from A to B — we take 100 meters as the amount of this maximal sinking of sea-level, as we deduced at length in a detailed study on the submarine topography of the Java-sea, the South China-sea and Malacca straits (Bibliogr. 17 and 19) — either simultaneous with it or shortly afterwards a rise of the land took place over an amount of about 30 meters, so that at present, after the rise of sea-level over 100 meters, we find the sea-level at C. Because there are no young reef limestones to be found in the coastal plain of Makassar we think it probable that the gradationplane has acquired its shape as such at the time of the negative shift of the strandline at the end of the Pleistocene. 2. — So the present coral reefs on the Spermonde shelf can for the greater part only have originated since the end of the Pleistocene; only those at the outer border of the shelf may have existed even during the Pleistocene; they gradually grew up as a barrierreef, now rising up from a depth of 30 fathom and showing many sandy islets on its crest. 3 and 4. — The Spermonde archipelago consists of a great number of submarine reefs, patchreefs and cays, in the most different stages of development in a manner quite analogous to those we described from the bay of Batavia; they also show the same distribution of the fauna in five different ecological facies. We visited this region at the end of the East monsoon; its influence clearly manifested itself in the shape of the islands; especially the small sandy islets still uncovered by vegetation were crescent shaped, the convex side turned to the East, resembling a sanddune or barchan. It is of much interest to state that the East monsoon obviously was not able to cause an analogous result on the shingle walls. These shingle walls showed a quite opposite orientation. They were all found at the west side of the sandy islands, their convexity turned to the West; on the outer slopes the reef flourishes, on the inner side we find the shallow reef flat (see fig. 17 and Pl. IV in Liter. 16); the sandy islands allways form the most eastern part and as no shingle wall occurs on this side and no coral growth worth mentioning can flourish (because of the large quantities of coral sand sedimentating there) the eastern side of the islands allways offers a good landing place. So, the geological structure of the islands as well as the distribution of animal life on the reefs clearly depends on a preponderating influence of the Westmonsoon, which blows unhampered on these islands across the Java sea and Makassar straits. The force of the east monsoon on the contrary, when blowing across the Flores sea, is broken, at least in the lower layers of the athmosphere, by the high mountains of South-Celebes (which form a shelter for the Spermonde archipelago against the force of the East monsoon). The small influence of the East monsoon is shown by the alternating shape of the sandy islets; its strength is not great enough to cause a permanent result in the orientation of the coralreefs, the shingle walls or the sandy islands. In comparing data on wind frequency, wind velocity and the wind effect in the isle of Noordwachter (Thousand Islands, Java sea; Lit. 17) and the bay of Batavia, we pointed out that data on the wind effect always give reliable values for geological conclusions; we showed, however, that data on wind frequency only, may sometimes give indications as to the probable wind effect but only when the wind in question is not hampered in its route by mountains (see fig. 5 and 6). Figure 7 and 8 show typical instances; most of the islands in the Spermonde archipelago are built in an analogous manner. Some islets, however, show some differences from this general type; e.g. the island Sarappo has no reefflat and no shingle walls; it seems to me that this is clearly explained by the fact that on its western side the Salisireef occurs (fig. 10); Salisi shelters Sarappo against the strong surf of the west monsoon. 5. — When discussing the coral reefs in the bay of Batavia I pointed out the curious fact that most of the islets show negative shift of the strandline of one to two meters, indicated by parts of raised coral reefs, and as the isle of Leiden formed an exception I said (Lit. 5, pag. 8 and Lit. 16, pag. 64): „The fact, that in some islands any sign of up heavel is wanting, pleads against an application of Daly’s „worldwide sinking of ocean level” to explain the facts in Batavia bay”. Since then on one spot of Leiden island a raised coralreef was also revealed on a stormy day in 1929. So my objection against a possible explanation according to Daly’s theory is of no value any more. This paper is the third in a series cited here below in the bibliography; in the near future a note on Sluiter’s coral reef at Krakatoa, a study on the coralreefs of Emmahaven (W.-Sumatra) and on the barrier reefs and atolls of the Togian islands (N. Celebes) will be published.