So much for observations on the flight of wild birds. I shall now briefly record some of the more accurate observations on the rate of flight of carrier pigeons.

Tegetmeier declares (Field, 22, i, 87) that the average speed of carrier pigeons is 36 miles per hour, whilst on two occasions a speed of 55 miles per hour was maintained for four hours in succession.

From experiments carried out in a covered gallery (Field, 1887, p. 242) it was shown that a pigeon flew at 33.8 miles per hour, whilst in the open another flew at 27.9 miles per hour.

In the Homing Fancier's Annual of 1892 it was recorded that in covering 82 miles in good weather a bird maintained just over 71 miles per hour. From the Scilly Islands to Wiltshire (215 miles) a bird kept up a speed of 50 miles per hour. In 170 miles a bird made 54 miles per hour, and in 104 miles it made 57 miles per hour. In a race from Banff to Hampshire a bird maintained 62 miles per hour in very favorable weather. Finally, a celebrated bird called "Volonel" on two occasions maintained over 60 miles per hour.

Doubtless other figures have been published, but I have been unable to trace them. From the data available it appears that the normal velocity of a carrier pigeon is from 30-36 miles per hour, but that when "homing" they can attain up to 60 miles per hour or over. Again arises the question as to whether migrants can accelerate their speed when actually migrating, in the same manner that a "homing" pigeon can hurry on its way when "homing." For reasons already given, I do not think they do, and there is certainly no evidence which even suggests it. The cases of rooks in the above table were certainly those of migrating birds, and indicate no hurry. The Rossitten birds were all on passage, and show no excessive speed. In fact, the only excessive speeds we have in the table are those of the two Lammergeier which were taken under abnormal conditions, the golden plovers which were escaping pursuit, and the Roubaix swallow. It is remarkable that this bird was also "homing," which may account for such an abnormal speed.

But swallows are most deceptive birds as regards their flight. They are in reality neither strong nor rapid fliers, and personally I do not attach too much reliance in the data of the Roubaix swallow. I do not believe any swallow is capable of anything approaching that speed unless assisted by a tail wind of 30 or 40 miles an hour, which, as is well known, is a hateful condition to a traveling bird.

The case of the Mosul swifts is interesting. The birds were probably not on passage, but simply feeding. It is known that swifts travel great distances in search of food and ascend great altitudes. In the Middle Atlas of Morocco, in the Himalayas, in Crete and Palestine, 4,000 or 5,000 feet and 50 miles or so in distance seems nothing to these incomparable fliers. I have had splendid opportunities of observing both the Alpine, common, and spinetailed (Chatura) swifts, and it has been a great disappointment to me that I have never been able to get a satisfactory estimate of their rate of flight, as they never continue on an even course. On a small island off the coast of Crete, I was recently given a good exhibition of what an Alpine swift can do. I was watching some of these birds feeding round cliffs in which several pairs of Eleonora's falcon were about to breed. Now, this delightful falcon is no mean flier, and as these swifts passed their cliff, the falcons would come out against them like rockets. The swifts would accelerate, and seemed to be out of sight before the falcons were well on their way. So confident were the swifts in their superior speed, that every time they circled round the island they never failed to "draw" the falcons, and seemed to be playing with them. I may add that these same falcons have little difficulty in overhauling and striking a rock pigeon-itself no mean performer. I have also seen on record the case of falcons and swifts somewhere in India, when the former failed time after time to come up with his quarry. I unfortunately can not trace the reference.

I hesitate to even guess at the speed to which a swift can attain when the necessity arises, but the main point is that this, the fastest of birds, can increase his "feeding" speed of, say, 70 miles per hour to a velocity which must exceed 100 miles per hour. There is little doubt that the speed of the golden plover in the table is an accelerated speed. Pilots in Mesopotamia have told me that whereas geese can not to any great extent accelerate, duck, when pressed, could attain a speed of about 60 miles per hour.

To conclude, I find that birds have two speeds-a normal rate which is used for every-day purposes and also for migration, and an accelerated speed which is used for protection or pursuit, and which in some cases nearly doubles the rate of their normal speed. Some of the heavier birds can probably only accelerate to a slight

extent. In this conclusion I am naturally excepting "courtship" flight, which is usually of an accelerated nature.

I also find, after eliminating abnormal conditions and observations based on meager evidence, that the normal and migratory rate of flight in miles per hour is as follows:

A BOTANICAL RECONNAISSANCE IN SOUTHEASTERN

ASIA.

By A. S. HITCHCOCK.

[With 11 plates.]

The following sketch of the region is based upon a visit made in 1921 to the Philippines, Japan, China, and Indo-China for the purpose of studying and collecting grasses for the United States Department of Agriculture.

The itinerary was as follows, the localities referring to places where collections were made:

Philippines: Vicinity of Manila, Los Baños, Baguio.

Japan: Yokohama, Tokio, Lake Hakone, Mt. Fuji, Nikko and Lake Chuzenji, Kyoto, Nagasaki.

China: Shanghai, Nanking, Kuling, Hongkong, Canton, Yingtak, Shiuchow, Whampoa, Lohfau, Macao, Pakhoi, island of Hainan. Indo-China: Haiphong, Hanoi, Vinh, Hue, Tourane.

PHILIPPINES.

These islands have been so well exploited biologically by the Bureau of Science that only time enough was spent here to obtain a general view of the grass flora. The native vegetation has been largely replaced in the vicinity of Manila by introduced species, so that most of the grasses here belong to the category of weeds.

The grass supply for horses and other animals in Manila is furnished partly by roadside species but largely by Homalocenchrus hexandrus. This species is found in wet soil in the tropics of both hemispheres, but it was surprising to find it cultivated for forage. Large fields in the vicinity of Manila are devoted to this grass, the plants being set out from divisions of the tufts, and treated much in the same manner as rice. The grass is cut, tied in bundles, and transported green to the city. The usual name applied to it is zacate, the Spanish name for grass.

One of the commonest grasses is cogon (Imperata cylindrica), which is found also in tropical Asia. It is an aggressive species,

101257-23-25

373