If we consider the regions in which cyclones appear, the paths they follow, and the direction in which they whirl, we shall be able to form an opinion as to their origin. In the open Pacific Ocean (as its name, indeed, implies) storms are uncommon; they are infrequent also in the South Atlantic and South Indian Oceans. Around Cape Horn and the Cape of Good Hope heavy storms prevail, but they are not cyclonic, nor are they equal in fury and frequency, Maury tells us, to the true tornado. Along the equator, and for several degrees on either side of it, cyclones are also unknown. If we turn to a map in which ocean-currents are laid down, we shall see that in every ‘cyclone region’ there is a strongly marked current, and that each current follows closely the track which I have denominated the storm-⊂. In the North Atlantic we have the great Gulf Stream, which sweeps from equatorial regions into the Gulf of Mexico, and thence across the Atlantic to the shores of Western Europe. In the South Indian Ocean there is the ‘south equatorial current,’ which sweeps past Mauritius and Bourbon, and thence returns towards the east. In the Chinese Sea there is the north equatorial current, which sweeps round the East Indian Archipelago, and then merges into the Japanese current. There is also the current in the Bay of Bengal, flowing through the region in which, as we have seen, cyclones are commonly met with. There are other sea-currents besides these which yet breed no cyclones. But I may notice two peculiarities in the currents I have named. They all flow from equatorial to temperate regions, and, secondly, they are all ‘horse-shoe currents.’ So far as I am aware, there is but one other current which presents both these peculiarities—namely, the great Australian current between New Zealand and the eastern shores of Australia. I have not yet met with any record of cyclones occurring over the Australian current, but heavy storms are known to prevail in that region, and I believe that when these storms have been studied as closely as the storms in better-known regions, they will be found to present the true cyclonic character. Now, if we inquire why an ocean current travelling from the equator should be a ‘storm-breeder,’ we shall find a ready answer. Such a current, carrying the warmth of intertropical regions to the temperate zones, produces, in the first place, by the mere difference of temperature, important atmospheric disturbances. The difference is so great, that Franklin suggested the use of the thermometer in the North Atlantic Ocean as a ready means of determining the longitude, since the position of the Gulf Stream at any given season is almost constant.