sonably say that the 94 stone-falls, about which we know comparatively little, seem decidedly to follow the same laws as the 116 falls about which we know so much more.

This conclusion is greatly strengthened if we take account of the effect of the earth's attraction in carrying the meteor's quit toward the zenith. Any stone must be moving downward when it enters the air. But the earth's attraction must change the direction of its motion during the approach to the earth. Hence the region of the heavens from which a stone can approach the earth is not bounded by the actual horizon, but by a curve which may be treated as a depressed horizon. This depression of the horizon is far greater toward the quit than toward the goal side of the horizon. The maximum depression for a stone moving in a parabolic orbit is about 17°. It hence follows that when the zenith is more than 73° and less than 90° from G, both the points G and Q are above the depressed horizon, and therefore that the 14 falls whose zeniths are between these limits, that is, are between the circles AA and PEPS, figure 2, should be left out of the count. The corresponding region on the Q-hemisphere is less than one degree in breadth, and contains one zenith point. We have left only 20 falls when the earth's goal alone was above the depressed horizon to be compared with 59 falls when the earth's quit alone was above the depressed horizon.

Of the 50 observed falls constituting the third group, of which the hour of fall is not stated, very few particulars other than the fact of fall are known. Although we are left without the power of saying that they indicate the same law as the other 210 falls, we find at the same time no reason to suspect the contrary. It is not unreasonable to assume that the wellobserved stone-falls are good representations of the whole group, and to affirm the three propositions with which I set out as true, in general, not only for the 210 stone-falls of the first two groups, but for the whole 260 stone-falls which are represented by stones in our cabinets, and in which the stones were seen or known to fall.

It also seems a natural and proper corollary to these propositions (unless it shall appear that stones meeting the earth are destroyed in the air), that the larger meteorites moving in our solar system are allied much more closely with the group of comets of short period than with the comets whose orbits are nearly parabolic. All the known comets of shorter periods than 33 years move about the sun in direct orbits that have moderate inclinations to the ecliptic. On the contrary, of the nearly parabolic cometic orbits that are known only a small proportion of the whole number have small inclinations with direct motion.

It also follows that in future reductions of these stone-fall observations it will be better to assume that the velocity of the stone in its orbit was not that velocity which corresponds to a parabolic orbit, but that which corresponds to the mean orbit of the comets of short period. The largeness of the perihelion distances has an evident bearing also upon the idea that these stones form the fuel of the sun.

The presentation of the argument here made has been incomplete in that the details of the investigation of individual stonefalls has been entirely omitted. Some of the determinations of the paths are, I think, as complete as I can hope to make them. But others must be regarded as provisional, since I hope to secure respecting them additional data. I hope at some future time to give a complete discussion of all these observed stone-falls. In the past I have been greatly indebted to friends for aid in collecting accounts of the falls, and I heartily thank them therefor. I shall be very grateful also in the future for unpublished observations of the stone-falls, as well as for observations that have been so published as not to be likely to have attracted attention. I bespeak the kindly aid of any who have made or have collected such observations.

ART. II.-History of Changes in the Mt. Loa Craters; by JAMES D. DANA. With Plates I, II, III.

[Continued from page 289; and also from xxxiii, 433, xxxiv, 81, 349, xxxv, 15, 181.]

SUPPLEMENT TO THE HISTORY OF KILAUEA.-Since the debriscones of Halema'uma'u, the great lava-lake of Kilauea, have a constitution and history unlike anything thus far reported from other volcanic regions, I add to the previous notes the following from a recent letter of Mr. J. H. Maby, of the Volcano House, dated March 8th. Mr. Maby writes that the cone has been rising since August, of 1887, until now the summit is "on a line with the outside walls of the crater, looking from the Volcano House." No additions have been made to the exterior, but instead, the eastern side (which Plate 5 in the last volume, from a photograph, showed to be in process of separation from the rest) "has slipped down a little and changed considerably its shape. Moreover the bottom or floor of the Great Lake with its lavas, is now within 40 or 50 feet of the top, which implies a rise of 30 or 40 feet in the same interval. The fires have been very active, and are now visible, from the house, on the east side of the cone; and the lavas on that side have flowed over into the deserted basin of New Lake, filling its lower portion. The lake on the west side of the cone has also much increased in size, being now nearly 300 feet in diameter; and it has thereby encroached on the doomed cone.

་

II. MOKUAWEOWEO, THE SUMMIT CRATER OF MT. LOA.

Maps.-A map of the island of Hawaii, reduced from the Government map, is here introduced (Plate 1) for the better illustration of the facts and discussions beyond.* It shows the topographic simplicity of the island-a fact not expressed in most of the small published maps, which generally (like that of the eleventh volume of the new Encyclopedia Brittanica) put in mountain ranges or ridges that do not exist. The map will enable the reader to appreciate the relative position of Kilauea and the Mt. Loa crater, their relative heights, the absence of water-courses from all of the mountain slopes except a small windward region; the large size of the valleys of the Kohala Mountains to the north; the positions of the great lava streams of 1840 to 1887; the routes of the roads (mostly bridle paths); the two routes to Kilauea, one of thirtyone miles on horseback from Hilo, the other of about half this distance, from Keauhou, the upper half a good carriage road; and also the districts into which the island is divided, and the positions of the principal villages.

The present form of the summit crater, Mokuaweoweo, is shown on the map by J. M. Alexander, Plate 2, reduced from the results of his survey. The height of the highest point, given on it, 13,675 feet, differs eighty-five feet from Wilkes's determination of the same point in 1841.

The history of the summit crater is mostly a history of the results of its eruptions, for few facts have been observed about the action within the crater. It has excited attention when an eruption has been in progress; but the chief outflows have begun below the summit and the source of the outflow is usually the only place reached. Still there is much to be gathered from the reported facts. My personal investigations have been confined to the base of the mountain, and the review beyond is hence almost solely from the accounts of others.

HISTORY OF THE ERUPTIONS FROM 1832 TO 1888.

1832, June 20.-On the 20th of June, 1832, according to Rev. Joseph Goodrich, lavas were discharged from several vents about the summit. The fires continued to be visible for two or three weeks, and were seen from Lahaina, 100 miles to the northwest. Nothing is known of any large discharge of lavas, and no mention is made of accompanying earthquakes.

*The Government map as stated upon it, is only a preliminary map, part of the survey being still incomplete.

+ Goodrich, this Journ., xxv, 201, 1834, letter of Nov. 17, 1832.

The outbreak of Kilauea in 1832 occurred about the same time, but possibly a few months later (xxxiii, 445, 1887 and XXXV, 15, 1888).

1834, January 29th.-Mr. David Douglas, the naturalist, who was the first to ascend Mt. Loa, describes the crater, in his Journal,* as having great chasms in the bottom that he could not fathom "with a good glass and the air clear of smoke" and says further: that "the depth to the bottom on the east side was by an accurate measurement with a pine and plummet, 1270 feet" that the southern part of the crater, "where the outlet of the lava had evidently been, must have enjoyed a long period of repose." He mentions hearing light hissing sounds from fissures in the summit that might "perhaps be owing to some great internal fire escaping." He adds, "There is little to arrest the eye of the naturalist over the great portion of this huge dome, which is a gigantic mass of slag, scoriæ and ashes."

1841, January.-Captain Wilkes was at the summit during the latter part of January, 1841.+ Lieutenant Eld, by taking angles from the bottom of the crater, made the western wall 784 feet high, and the eastern, 470 feet. The only sign of activity was the escape of steam and sulphur gases from many deep fissures over the bottom, especially on the west side. The fissures had generally a N.N.E-S.S.W. direction. There was one cinder or scoria cone at the bottom, according to Dr. Judd, toward the southwest side, having a height of about 200 feet.. Other steam cracks were observed outside about the pit crater of the south-southwest end; and one, which they designated the great steam-crack, led from the top of the mountain a long distance down its sides toward the south;" and a great depth was indicated by the reverberations from a block of lava which was dropped into it. Small driblets of lava were observed along some of these fissures; indicating feeble ejections at the very summit. In Wilkes's map, as shown in the outline copy on the next page, seven small cones are faintly represented on the bottom of the crater, although the descriptions speak of only one.

66

1843, January.-In January of 1843 began one of the great outflows. It continued for about six weeks. Clouds above on the 9th made the first announcement to the people of the

* Companion of the Bot. Mag., ii, 175, 1836, and in a letter to Capt. Sabine, dated May 3, 1834, Journ. Geogr. Soc., iv. 333, 1834. See this Jour., xxxiii, 436, 437, 1887, on the letter to Dr. Hooker and the evidence against it.

Narrative of the Exped., iv, 152, 156, 159. The descriptions of the crater are from descents made into it by Dr. Judd of Honolulu (on p. 152) and Lieut. Henry Eld (on p. 156). Wilkes's map has its longer diameter, through some mistake, north-and-south in direction.

islands. During the following night, according to Dr. L. Andrews,* a brilliant light appeared at the summit, looking, as Mr. Coan states, like "a small beacon fire."+ In a week

The Summit crater, after Wilkes, January, 1841.

the light disappeared. In the mean time the lavas had commenced their discharge. Mr. Coan ascended to the source, about 13,000 feet up, and found two large craters near together, very deep and active. The source given on the map is at least 2000 feet lower. The stream of lava flowed toward Mt. Kea, but gave off a westward branch, toward Hualalai, near its source. At the base of Mt. Kea, a branch went northward toward Waimea, and another eastward toward Hilo. Mr.

* Andrews, Missionary Herald, xxxix, 381, letter of Feb. 6, 1843.

† Coan, ibid., xxxix, 463, letter of Feb. 20, 1843; xl, 44, letter of April 5; this Journal, II, xxvii, 411, 1859; Life in Hawaii, 1882, p. 270.

AM. JOUR. SCI.—THIRD SERIES, VOL. XXXVI, No. 211.-JULY, 1888.