a northern bed it occupied 550 years before, and joining the Ta tsing ho, discharged into the Gulf of Chihli, several hundred miles further up the coast. In 1887 a terrible inundation occurred by the river bank giving in, and towns and villages were swept away.

To hinder its overflowing, embankments hem it in, some nearer, others farther, ranging one behind another at variable distances. In this manner, if one gives way, another prevents the inundation. In its present state, the work is still very inefficient, the dikes being weak, and constructed with materials that offer insufficient resistance.

The mud and sand which frequently obstruct the Yellow River, render it also very difficult of navigation. The only portion where it can be availed of, is to the north of Honan, and in the last 25 miles of its course. But even in this part, a shoal prevents junks except of very light draft from passing.

The flow of the Yellow River varies much with the season. It has been reckoned to be a little over 4000 cubic yards per second, in its middle portion, near Tsinan Fu(Shantung). The flow is three miles greater in the flood season. It is on the whole relatively small for such a great river, but this is partly accounted for by the waste of the water that filters through the embankments. The mud and sand, which it unceasingly deposits in the Gulf of Chihli, constantly lessen the depth of the latter, and form there new alluvial lands. Opposite the former mouth of the river (1851) one can see what great quantity of sediment was carried in its waters.

The last serious breach in its dykes occurred in September, 1902. The Chinese engineers showed great ingenuity in effecting its repair. The breach was near Liu-wang-chuang and was 1500 yards, through which most of the river flowed. It was repaired by building out from each side, dams in the form of a series of pakwerks of kaoliang stalks and sacks of clay, each pakwerk or buttress being joined to the previous one by ropes and piles.

Kaoliang is a kind of sorghum, probably identical with Barbados millet. The core of the stalk, except for a very thin and weak covering, is entirely pith, but it has a matted

bunch of fairly hard and strong roots which form its chief virtue for construction work. The stalk is about 6 feet long, inch in diameter, and the bunch of roots, 3 to 5 inches in diameter. The face of the werk including the sides is composed of the roots which mat together and make a splendid surface for keeping out water.

The final opening in this breach of 1500 yards, after pakwerks were built out from each side, was reduced to 55 feet, and this, after two disastrous attempts in which the lives of many workmen were lost, was effectively closed on March 16, 1903, by letting down a huge mattress of kaoliang stalks and sacks of clay, the mattress being anchored to the side of the river by a great many 15-inch hawsers so as to prevent canting due to impact of current. Over one hundred 8-inch ropes spaced closely together were stretched across and belayed to anchor piles. On these were than placed in alternate layers the kaoliang stalks and sacks of clay. When these materials reached the level of the sides of the dam, the ropes were manned, and at a given signal were each lowered 1 foot on each side.

The rush through the opening was reduced by the construction of a deflecting groyne on the up river side of the breach, constructed similar to the pakwerk, and projecting some 120 feet into the current. The width of the river channel abreast of the breach had been 600 feet but was reduced to 300 feet by the formation of a sand bank on the opposite side of the river.

The control of the Yellow River is today one of the most pressing of China's physical problems. Experience has shown that the diking of such rivers is insufficient and almost futile. Captain William Tyler, coast inspector of the Chinese light house service, has presented a report on the Yellow River published by the inspectorate-general of customs at Shanghai in 1906, in which he proposes to control the river's lower reaches by providing for the depositing of the silt by deliberate flooding of large areas along the river, that is, to regulate its floods.

For this as well as for other rivers subject to floods, very comprehensive surveys should be made and the history of

the streams involved secured as accurately as possible. It has been surprising to me to see how often the railway engineers have erred in China, from their inadequate attention to this factor of river history, and after costly experience have had to redetermine the level of their tracks when flanking a river or lake or to rebuild their bridges crossing a stream.

Another example of Chinese engineering skill employed in protecting the land from the "misbehavior" of rivers is the great sea wall along the north side of Hangchow Bay, of which I have elsewhere given a full account, and is, considering the difficulties under which it was built, one of the finest pieces of such engineering to be found anywhere.

3

The Hangchow bore is caused chiefly by the funnel-shaped character of Hangchow Bay, combined with extensive mud and sand bars that occupy its mouth so that the tide in coming up the by instead of gradually rising, banks up near the entrance. The difference in level is such that a great wall of water anywhere from 10 to 30 feet in height rushes up the bay and up the river at a speed which sometimes equals that of an ordinary express train. This occurs twice a day and always amounts to a considerable wave, although sometimes it is much larger than others. Of course, an exceedingly strong wall is required to keep out such a mass of rushing water from the surrounding country which, as a matter of fact, is in many places below the level of the river. Some 750,000 taels are spent each year for the upkeep of this wall. It is built of heavy granite blocks joined together with double iron wedges and besides the wall which is ordinarily 30 feet in height, there are two granite platforms or ledges each edged with a multiple series of long piles driven into the sand, constituting one of the strongest sea footings that could be secured to keep the wall from being undermined. At intervals of about half a mile for a good part of the wall, there are pakwerk buffers to deflect the current of the bore.

As if these two cases of the devastating forces to be overcome were not enough to develop resourcefulness and hardihood in the people compelled to face them, the coast of

Popular Science Monthly, February and March, 1908.

China is subject also to frequent typhoons, many of which are destructive of life and property in the extreme. I shall attempt no description of the typhoon and its origin, but wish by my pictures to call attention to the tremendous destruction caused by such storms.

In order that there may be more timely warnings, there are needed more observatories and better coördination in the work of existing observatories throughout the Orient. There are, at the present time, well-established observatories at Zikawei near Shanghai, maintained by French Jesuit missionaries; at Hong Kong, maintained by the British colonial government, and at Tsintau, the German concession on Shantung promontory. These observatories are more or less in receipt of communications from the observatories at Manila and Tokio, and there are also observations of more or less regularity at various light house stations along the China coast and at some of the ports by the harbor masters in the customs service. But, there is a great deal more than this to be done, and the whole work needs to be put upon a sound basis in its scientific work and in its administration.

One of the best things that Sir Robert Hart did in connection with the customs service was to give the China coast its needed light houses, so that today it has a chain of such that will rival those found anywhere.

The country at large needs also a weather service. While this is primarily the government's duty, missionary colleges at the present time have a real opportunity to assist China in this connection. The physics department of each of the colleges throughout the land should make adequate and systematic meteorological observations so that when the time comes when the government is able to organize a service on its own basis, there will be qualified observers available and an accumuation of valuable data upon which valid generalizations as to the meteorological forces in China may be based. All this is closely connected with re-forestation, extensive farming and the control of rivers. There should be a thorough meteorological survey as a necessary preliminary if these problems are to be adequately solved, for it will require at

least twenty years to gather the data that will render generalization valuable.

Partly as a preliminary to this, the Carnegie Institution of Washington has for over six years been carrying on magnetic observations throughout China in accordance with plans which I submitted to them to be carried out in connection with their magnetic survey of the North Pacific, and I shall treat briefly of the aims, scope and results of this magnetic survey as a distinct contribution to the solution of China's physical problems. The results of such a survey are necessary to the land surveyor and to the navigator in order that when a magnetic compass is used either to steer a ship at sea or to run the lines of a survey on land, the user may know the amount accurately by which the needle deviates from the astronomical or true north. At each station, the observer determines latitude and longitude by astronomical observations; the compass deviation or declination, the dip of the magnetic needle and the intensity of the earth's magnetic force at that place. All these elements are necessary in order to predict the way in which the magnetic declination from the north will vary with the years. The Carnegie Institution of Washington has extended its operations to China as a part of its plan to supplement the work of the constituted governments, who have not yet organized scientific services. Already fifteen of the provinces have been traversed and about a hundred stations have been occupied at intervals varying from 25 to 100 miles. The results for the years 1905-1910 are just being published and had previously in part been made available to those chiefly interested. Only three of the most western provinces and the greater part of Mongolia remain to be covered in this preliminary reconnaissance.

Connected with this matter of surveys is the whole problem of reform in weights and measures in China which are in utter confusion today. But while really a part of your topic, I shall not attempt here any discussion of this item.

During my survey trips I have of course come to appreciate very feelingly the problem that China has with reference to