it is therefore dangerous to lie-to in a typhoon, particularly before you are sure that the centre is past. Vessels near the coast of China, or in the Formosa Channel, generally seek refuge in the nearest typhoon harbour indicated in the Directory.

The wind shifts faster the nearer the centre you are. If the barometer falls rapidly and the wind does not change its direction, and when the gusts continue to increase in force, your ship is in danger of entering into the central calm of the depression with its mountainous and confused seas, which is by all means to be avoided, as it is the high cross seas that do the most damage, and not the force of the wind. When once you are caught in a typhoon you should make no sail, except what may be necessary to steady the ship, till the gusts continue to decrease in force and the barometer has risen for some time. Very deceitful lulls are reported to occur during the raging of a typhoon. The master of a sailing-vessel is said to have put up topgallant sails after getting into the central calm. Of course he could have had no reliable barometer on board.

In storms encountered in higher latitudes, where the incurvature of the wind is not so great as in a tropical hurricane, the right-hand semicircle is termed the dangerous semicircle, as a ship running before the wind is in more danger of crossing the path of the storm in front of the centre and perhaps be overtaken by it; but in a typhoon there is not much to choose between the semicircles. A dismasted ship is in danger of being carried into the centre from any quarter.

However, the right-hand semicircle is also in a typhoon generally more dangerous than the other, both with regard to the risk of crossing the path in front of the centre, and also, as remarked above, with regard to the force of the wind and consequent greater sea disturbance. A ship experiencing a northerly gale and a falling barometer in the China Sea in the typhoon season is generally in greater danger than another experiencing a south-westerly gale.

When you have ascertained in which semicircle your vessel is situated, you should, if in the right-hand semicircle, keep the wind as long as possible on the starboard tack; and if in the lefthand semicircle, you should run on the starboard tack, or heaveto on the port tack, so as to let the ship come up as the wind backs and run no risk of being taken aback. As explained further on, a typhoon encountered in a low latitude moves so slowly that a steamer or fast sailing-ship has a fair chance of running away from it, but farther north, when the centre proceeds at the rate of thirty or forty miles an hour, it requires careful management even supposing you have ample sea-room.

Typhoons are dangerous on the open sea, but they are still more to be feared in open anchorages or near lee shores. Along the south-west coast of Formosa and elsewhere, a ship must in the south-west monsoon be prepared to run to sea at very short notice, as in some of the harbours you could not lie with any chance of riding out a typhoon. A steamer at anchor should get up steam as soon as the wind rises above the force of a strong breeze, and a sailing-vessel should take down the top-masts. The direction in which the wind is going to shift must early be determined so as to select a sheltered anchorage. If the centre passes very near the anchorage, the berth may have to be changed: to the other shore during the lull, before the wind shifts to the opposite quarter.

A ship moored by a single anchor with her head to the wind, will swing with the sun in the right-hand semicircle and against the sun in the left-hand semicircle. If two anchors are dropped, the anchor on the advancing bow should be let go first, therefore a ship in the right-hand semicircle of a typhoon should first drop the port anchor and afterwards the starboard, in order that she may ride with open hawse. And a ship in the left-hand semicircle should first drop the starboard anchor. But ships have to ride with a long scope in a typhoon, and as they are liable to drag the anchors, some prefer to drop the second

anchor to veer upon if the first should not hold,

11. The force of the wind and the appearance of the sky do not always furnish a reliable guide to determine how far you are from the centre of a typhoon. The dimensions are different in different typhoons, and near land the strong winds are often so irregularly distributed that in a place near the centre less wind may actually be experienced than at some distance farther away from it. Also the 11-point rule for ascertaining the bearing of the centre fails near some shores if the centre is not near at hand; thus there often blows a steady easterly gale along the

Sea, and the gale blows often steady from north-east about the northern entrance to the Formosa Channel when there is a typhoon in a more southern latitude.

The surest of all warnings is furnished by the standard barometer on shore and the compensated aneroid on board ship; you are all right if you can put your vessel on the tack that will keep your barometer rising. But in order to understand the indications of the barometer you will have to keep a regular meteorological register. The master of a vessel who does not look at his aneroid till he is in a typhoon, does not derive half the benefits from his observations that he would have enjoyed had he watched it beforehand. He might perhaps have avoided the weather he is now experiencing, or even have benefited by the favourable winds and sailed round the typhoon. No doubt the time is approaching when underwriters will stipulate that the indications of an aneroid or a marine barometer must be regularly registered on board a vessel insured by them.

On the other hand, it would not be fair to ask the mariners to keep complete meteorological records, such as are kept in the lighthouses out here. Some seamen have a taste for this kind of work and make very useful and fairly accurate observations, but, for instance, the readings of dry and damp bulb thermometers taken on many vessels are of very little use.

The tube of the marine barometer has to be so much contracted to stand the incessant pumping and danger of breakage, that the instrument is sluggish and often reads half an inch or more too high near the centre of a typhoon, Some cheap wooden barometers cannot be registered below a certain height, the cistern being too small to hold the mercury that comes out of the tube. Of course some cheap aneroids are no better, and even a first-class compensated instrument requires to be tho roughly verified, as the scale is never quite correct, but they act nearly as quickly as first-class standard barometers, and for use on board ship the instrument that is quickest in its indications must be preferred. The objection to the use of the aneroid is founded on the fact that its index-correction changes gradually; but then this can be determined and allowed for by reading it off as often as the vessel enters a port, such as Hong Kong, where correct meteorological observations are constantly being

made.

The best hours for making observations are 4 a.m., 8 a.m., &c., up to midnight inclusive. The observations should consist in readings of the aneroid, temperature (this is no use except when the thermometer is placed well forward so as to be espored to the wind, though in a position sheltered from the sun and the rain), direction and force (0-12) of the wind, direction whence coming of the clouds, amount (0-9) of sea-disturbance, an weather (Beaufort's notation). For further particulars the "In structions for making Meteorological Observations, prepared for use in China," published in 1883 by the writer, may be

consulted.

From 4 a m. to 10 a.m. the barometer is rising, falling from 10a.m. to 4 p.m., rising from 4 p.m. to 10 p.m., and falling from 10 p.m. to 4 a.m. It reads highest at to a.m. and lowest at 4 p.m. During the approach of a typhoon this regular daily variation may be masked, but it goes on all the same, and must be taken into account when the barometer begins to fall before a typhoon. Thus if it has fallen a certain amount between 10 a.m. and 4 p.m. you must subtract the normal descent be tween these hours in order to know how much of the fall is due to the approach of the typhoon, and if it were between 4 p.. and 10 p.m. that it fell, you must add the normal rise for the same purpose.

and to sea-level, does not fall below 28 80 inches. In others t In many typhoons, the barometer, reduced to 32° Fahrenheit, falls as low as 28.50. Lower readings of the barometer appear to be rare, but it has been stated to have fallen much lower. The rate at which your barometer is falling depends upon your approach to the contre, and in consequence upon the rate st which this is travelling. For this reason it is not safe to dram conclusions concerning the amount of wind to be expected from the rate at which the barometer is falling, but to some extent, of course, this may be done. Remember, that, when the barometer has fallen to its lowest reading and begins to rise, you may expect to experience as much bad weather as you have already gone through.

does not experience any strong wind or bad weather till within a hundred miles or so of the centre, and as the typhoons are most violent in that locality, it is very important to look out for the premonitory signs referred to in the first section of this article, taking into account that the dimensions of a typhoon are so small there. On the other hand, they move at so slow a rate that you may run away from them if you are aware of the danger in time, so much more as you may be sure that a typhoon in that locality is directing its course to somewhere about westnorth-west or north-west, and most likely in the first-named direction. So it is better to get to the eastward of it. Nearer

the Philippine Archipelago the typhoons usually take a more northerly course, moving north-west or north-north-westward. But frequently they continue their course west-north-westward and cross the islands to enter the China Sea. In spring and autumn they have even been found to move westward and turn south-westward after entering the China Sea. But when you are east of the Philippines you should try to get your ship southeast of the typhoon by crossing the path behind the centre, if possible. If you are going northwards, you will then benefit by the fresh south or south-west breezes, taking care not to approach too near to the typhoon, whose progressive motion may not be more than six miles an hour. You will probably have squally and wet weather.

When all the paths of the typhoons that have been investigated by the writer in the course of the last three years are laid down on a map of the Far East, the picture looks much like a fan, the paths, with a few stray exceptions, radiating from the locality referred to above, and running in all directions between west and north, but most of them at first westward and then north-westward. In a higher latitude they generally recurve and pass off to the north-east, after first, of course, having turned northwards. Every typhoon does not recurve; in fact, as stated above, some of them finally disappear in the China Sea after turning to the south-west. The others recurve betwen 20 and 40° latitude, and 115 and 130° longitude. The Middle Dog Lighthouse is situated in the centre of the region of recurvature.

The normal path is therefore, roughly speaking, a parabola, whose axis lies from east to west, and whose apex is turned westward and lies within the region indicated above. But each path individually is anything but a regular parabola, and the deviations are evidently due to the influence of the coast-line of the mainland of Asia and to the mountainous islands (especially the high mountains of Formosa) as well as to the prevailing winds. For there is no doubt that the progression of a typhoon is the effect of the wind prevailing at the time, not necessarily at the surface of the earth, but at a somewhat higher level in the almosphere, which agrees with the direction of the clouds, that have, as explained in the first section of this article, been found to move nearly straight towards the centre in the posterior semicircle. If, however, the wind at the surface of the earth is strong, it is at times plainly seen to blow the typhoon before it. The typhoons do not appear to move south-westward in the China Sea except when the north-east monsoon is strong, and in the summer of 1885, when the south-west monsoon was strong, most of the typhoons moved northwards while yet to the east of Formosa. This is then the reason why the typhoons depend upon the season of the year. They are likewise deflected from their previous path when exposed to strong winds blowing out of open channels, such as the Formosa or Corea Channels, in which case the speed of their progress is sometimes abruptly increased.

The average rate of progress of the centre of a typhoon in 11 latitude is 5 miles an hour. In 13 it is 64, in 15 it is 8, in 20 it is 9, in 25 it is 11, in 30° it is 14, and in 324 latitude it is 17 miles an hour. The rate of progress does not vary perceptibly in case of typhoons south of 13 latitude, so it is well for masters of vessels to know this, but it is more variable the farther north you go. In 32 latitude it ranges between 6 and 36 miles an hour, so that you cannot at all be sure that a typhoon, which you may happen to be near, will travel at anything like the average rate of progress in that latitude.

In Observa ions and Researches made at the Hong Kong Observatory in the year 1884," the writer suggested the division of typhoons into four classes according to the paths which they usually follow. Of course abnormal instances, such as for instance are presented by the typhoons that originate in the China Sea, occur occasionally in China as well as elsewhere, but they are comparatively rare.

Typhoons of the first class occur at the beginning and the end

of the typhoon season. They cross the China Sea and travel either in a west-north-westerly direction from the neighbourhood of Luzon towards Tonquin, passing south of or crossing the Island of Hainan, or, if pressure is high over Annam, they travel first westward and subsequently south-westward. They can generally be followed for between five and six days.

Typhoons of the second class are the most frequently encoun tered, and their paths can be traced farthest. They generally travel north-westward while in the neighbourhood of Luzon, and either strike the coast of China south of the Formosa Channel, in which case they as a rule abruptly lose the character of a tropical hurricane, recurve in the interior of China, re-enter the sea somewhere between Shanghai and Chefoo (thereby regaining some of their past violence), pass across or near to Corea, and are finally lost sight of in their motion towards about castnorth-east; or they pass up through the Formosa Channel, recurve towards north-east, and pass along the coasts of Japan; or they may strike the coast of China north of Formosa. Typhoons following the latter path originate further east of the Philippines than the others. They either continue their motion north-westward, in which case they are soon lost, or recurve and pass north-eastward near Corea. Typhoons of the second class occur from June to September inclusive, but are most common in August and September. It appears that a third of the typhoons belong to this class. They can be followed on an average 7 days, or rather between 5 and 12 day-,

Typhoons of the third class are probably the most numerous of all, but are not encountered so frequently as the typhoons of the second class, and therefore the existence of a typhoon of this class is sometimes only suspected, although it of course influences the weather along the eastern coast of China through the fine weather area with which it is surrounded. They pass to the east of Formosa, travelling northwards. After recurving, they generally pass near Japan, but sometimes a typhoon of this class continues to move north-north-westward and does not recurve till west of Corea. They prevail in the same season as typhoons of the second class, and may be traced on an average during 7 days, or more correctly between 3 and 12 days. A typhoon of this class frequently follows after one of the second class. It is a well-known fact that depressions are attracted towards places which have just been traversed by a depression.

Typhoons of the fourth class pass south of Luzon, travelling westward, or first in this direction and then south-westward. They occur at the beginning and the end of the typhoon season, while the north-east monsoon is strong, namely in April and late in autumn, but are very rare. They are said to be more violent in autumn than in spring. Existing in so low a latitude, their dimensions are, of course, very limited. The writer has not been able to follow them for more than a day or two.

The number of typhoons that are known to have occurred in each month of the year, expressed in percentages of the total number of typhoons, is as follows:-January 2, February 0, March 2, April 2, May 5, June 5, July 10, August 19, September 27, Octo ber 16, November 8, and December 3, These figures prove that typhoons are most frequent during the month of September, but they also show that, strictly speaking, there scarcely exists a well-marked typhoon season. On an average there are 15 typhoons every year, but typhoons in different years exhibit

some variations.

IV. The writer on his arrival in the colony in 1883 found that meteorological observations were received from a few of the Treaty ports, &c., and were published in the local papers; and seeing that these returns would only have to be corrected and reduced, as well as slightly extended, in order to be of great value to the shipping, he took upon himself to effect this. Subsequently, as the official work of the Observatory was fully started, he would have had to give up this purpose had not the Government decided to support it. Thus originated the China Coast Meteorological Register, which is published daily from here. It contains, at present, observations of the principal meteorological elements, which are received through the cooperation of the great telegraph companies from Manilla, Bolinao (Luzon), Haiphong, Hong Kong, Amoy, Foochow, Shanghai Nagasaki, and Vladivostock, but the number of the station might with advantage be extended. It gives also information about the weather prevailing in the Far East, and more or les rongh intimations concerning such typhoons as happen to be indicated by the telegraphic returns, as well as by local observations. Subsequently more or less extensive monthly meteore

Many of the anchorages along the south-eastern coast of China and the south-western coast of Formosa afford excellent shelter against north-east winds, but would prove to be much worse than the open sea during a heavy southerly gale. If you observe a northerly gale and a falling barometer, by far the surest signs of an approaching typhoon, and appearances are rapidly getting worse, then occasions may possibly occur when you may be under the dire necessity of running southwards with the northerly gale, against the rules laid down by meteorologists, and bring your ship into a most dangerous position in front of the centre. you may happen to be better off there after gaining ample searoom, than in the snug anchorage, where your vessel would be smashed against the rocks as soon as she began to drag her anchors when the storm burst upon her from the south, although the south-west storms experienced along the south eastern coast of China during a typhoon that enters the mainland are as a rule less violent or protracted than the preceding storm from the north.

But

Suppose that after leaving Hong Kong bound for a northern port you were to ascertain the existence of a typhoon about to cross your course in front of your vessel, and you experience, say, a strong breeze from the west-north-west. If you do not alter your course, you may, from the fact of its subsequently appearing to be a hanging gale and seeing the mercury falling in the barometer, draw the erroneous conclusion that you are on the path of the centre of a typhoon coming down on you from the east-north-east. In such a case it does not appear to be advisable to scud before the wind, it being decidedly wiser to heave-to. Then if the gale is observed to begin to back towards south-west, you may run southwards and shape your course so as to sail round the typhoon. Masters of steamers leaving Hong Kong while the red drum is hoisted generally lose no time in running southwards as soon as the typhoon is observed to have taken a north-western course, and suffer very little delay in consequence.

Steamers bound for Shanghai are, while between Foochow and Ningpo, liable to experience the northerly gales that precede a typhoon of the second class travelling north-westward and about to strike the coast in that locality. Not wishing to expose their ships to the high cross seas round northern Formosa, the masters generally take them into the nearest typhoon harbour in order to wait till the centre has entered the mainland, and then run northwards with the southerly gale.

These few examples will be sufficient, the more so as the further consideration of the subject would lead into details with which the writer is not familiar, being possessed of no further knowledge of navigation than the little that can be gleaned from the inspection of log-books and from occasional conversations with masters of vessels of many years' standing. The writer has invariably found these gentlemen ready to recount their experi ences and to communicate any information, as soon as they found that it was required for scientific purposes exclusively. The master of a vessel, after encountering a severe typhoon, has often to undergo the vexation of seeing every manoeuvre of his subjected to the comments of those unaware of the hundreds of things he has to take into consideration besides the law of storms, and who were comfortably ensconced in their houses while he was experiencing the typhoon with its fierce gusts, interrupted by the, if possible, more ominous lulls, during which he cannot see three ship's lengths before him, the mountainous waves in which his good ship is but a "cock-boat," the loudest shouting inaudible, drowned in the roar of the tempest, boats and everything movable having been washed overboard, rudder gone, and perhaps one of the masts thumping at her bottom, while the seas threaten at every moment to swamp the ship.

VI. The origin of a typhoon is not quite understood, but appears to be connected with an abnormally high temperature and humidity in some place in comparison with the neighbour hood. Over such a place the hot air expands, ascends, and is thereby cooled. But the heat liberated by the consequent con densation of water-vapour retards the rate at which it cools on rising in the atmosphere, and enables it to rise still further. When the air has risen to a high level, it effects there an increase of barometric pressure, in consequence of which the upper air is set in motion out towards the circumference of the area in question. Thus a fall in the barometer at the surface of the ea in the middle of the hot and damp region is effected, and the surrounding air is impelled in towards the centre. The motions inwards at the surface of the sea, and outwards at a high atmospheric level, are, of course, contemporaneous, and one is assisted

by the other. But air in motion is deflected towards the right in the northern hemisphere, owing to the rotation of the earth, except at or very near the equator; whence typhoons do not exist in that locality, for if the wind continues to blow into the depression it is soon filled up. Owing to its deflection towards the right, the wind is caused to move in a curved path in towards the centre, and the centrifugal force, developed in the curvilinear motion, deflects it still further from the straight line leading into the centre. The friction between the wind and the greatly disturbed sea-surface likewise retards the entrance of the air at the sea-surface into the central parts of the depression. But the air at a higher level in the atmosphere is subject to little friction, and escapes therefore at a greater speed from the central highpressure area at that level. It is, therefore, apparent that once a cyclonal motion is started under circumstances favourable for its continuation, it tends to increase and to spread outwards.

There is, however, this important difference between a typhoon and a tornado, that the latter is taller than it is broad, whereas a typhoon perhaps does not reach above an altitude of four miles, while its horizontal diameter may amount to upwards of a thousand miles. Moreover, it is not at all likely that the centre at a higher level lies vertically above the central calm at the earth's surface, or even that the centres at different altitudes are situated in a straight line. We are, therefore, scarcely entitled to speak of an axis in a typhoon.

The spirals described by the air particles approaching the centre in a typhoon are known as logarithmic spirals, but unless a typhoon is stationary, which is perhaps never the case in Nature, new portions of air are constantly set in motion in front of the centre and others left behind by the typhoon.

As already remarked, the progressive motion of typhoons is evidently caused by the wind prevailing, if not at the surface of the earth, at any rate at a higher level. That the principal part of the disturbance is situated high above the surface of the earth is proved by the fact that the centres of typhoons pass across mountains several thousand feet high, and also by the circumstance that the difference between the temperature at the Hong Kong Observatory and at Victoria Peak is not perceptibly affected by the approach of a typhoon, for we cannot very well assume that the temperature of a vertical column of air is lower near the centre than outside the cyclone. The mountains referred to are situated on islands, and while crossing them the typhoon derives its store of water-vapour from the surrounding sea for as soon as the centre has entered the coast, and is on all sides surrounded by dry land, it ceases to exist as a tropical storm, and can only be traced in the registers through a slight fall in the barometer, a freshening of the wind, perhaps amounting to a moderate gale at stations crossed by the centre, and wet weather. Inland in China the bull's eye of a typhoon does not appear to be observed.

As the wind blows more straight into the centre the nearer the equator you are, it follows that more air enters the typhoon from the southern semicircle than from the northern, and this is one of the reasons why typhoons nearly always move in a northerly direction. Moreover, the difference increases together with the dimension of the typhoons, which explains why they expand and accelerate their progressive motion at the same time.

The foregoing observations contain the principal practical results of investigations of about forty typhoons, continued during a period of three years. The mariner into whose hands this article may fall is advised to determine for himself the direction in which the centre of a typhoon, which he is experiencing is travelling; for although typhoons of the classes enumerated are by far the most common, he never can be quite sure that he has not to do with an exceptional case, and quite possibly a case that is not found among the forty typhoons referred to above. By the time that we shall be in possession of full and trustworthy investigations of a couple of hundred typhoons, we may expect to have complete lists of the sub-classes of the four classes of typhoons, and to be better acquainted with cases of rare occur rence, for, after all, the typhoons are of a simpler construction and their paths more regular than is the case with storms in Europe. Typhoons are so violent near their centre that the whole disturbance is evidently ruled thereby, whereas storms in the North Atlantic and in Europe appear to be made up of a lot of local eddies, some of which are by degrees detached from the chief disturbance and form subsidiary depressions. The writer has not been able to ascertain the existence of a subsidiary depression in the China Seas during the last three years, and it is therefore doubtful whether such ever occur.