Need of thorough surveys. As has been stated, a good chart requires that a thorough and correct survey be first made of the region to be charted. It is said that men are very apt to accept as true anything they see on a map. As to the nautical chart the mariner is likely to be somewhat more critical, however, and it is well that he is. The difficulty of charting an invisible surface such as the bottom of the sea is great, and the proportion of the navigable waters surveyed in sufficient detail to be at all certain of the absence of uncharted dangers is small.

The planning of surveys in a new region, such, for instance, as the Philippine Islands, presents many interesting problems, on the solution of which the effectiveness in chart results and the cost of the work materially depend. Many local conditions must be taken into account. The surveys are made on opposite coasts according to the seasonal winds and rainfall. In some parts fair-sized steamers are necessary; in others launches and small boats can do the work more economically. Shore parties with land transportation are used for portions of the work where the country permits. Natives are employed as far as practicable for the classes of work they can do; the Filipinos, for instance, make good sailors on the vessels and excellent penmen in the office.

The following is a brief outline of the steps of a complete survey for charting purposes, according to the present practice of the United States Coast and Geodetic Survey. These are given in their logical order, though in actual work this order must often be departed from. In this Survey the methods of control have been of a high standard; that is, the main stations have been accurately determined and permanently marked and described, and this has proven an advantage in the joining together of the original surveys and resurveys.

Astronomical observations. To locate on the surface of the earth the area to be charted, astronomical observations are required for the latitude and longitude of one or more points. In the best practice the longitude of a point is obtained by observing the transits of stars to get the local time, and sending time signals by telegraph to obtain the difference from the local time of some other place whose longitude is known. The latitude is observed by measuring the difference of zenith distance of pairs of stars crossing the meridian north and south of the zenith. The azimuth or true direction of some line is also obtained from star observations, usually by observations with a theodolite on a circumpolar star. Much existing chart work depends on positions determined by less accurate methods, as, for instance, longitudes obtained by transporting chronometers between the known station and that to be determined, or by observations of moon culminations, and latitudes obtained by direct observations of the altitudes of stars with theodolite or sextant.

Triangulation. The main framework of the survey consists of a series of triangles connecting prominently located points which are permanently marked in the ground and the location described so that they can be found at a future time. At long intervals in the survey base lines are laid out and carefully measured with steel tape. Signals are erected over the points, including those at the ends of the base line, and angles are then measured at the various stations. From the measured length of the base and the angles the lengths of the sides of the triangles are computed, and from these lengths and the latitude and longitude of one point the latitudes and longitudes of all the other points are obtained. When several astronomically determined points are connected by such a triangulation a complication arises from what is known as "deflection of the plumb line," which is the angular amount by which the actual sea-level surface of the earth departs from the symmetrical figure of revolution, owing to the variations in the density of the earth's outer layers. The distance between two points as measured by triangulation thus differs from the distance computed from the astronomically determined positions. If this irregularity were not taken care of by adopting mean positions, the discrepancy in joining up different surveys would in extreme cases amount to about half a mile.

Survey sheets are next prepared, of suitable size and scale. On each sheet a projection is laid down, that is, the meridians and parallels are drawn, and all the points determined in the triangulation are plotted in their true relation. Usually separate sheets are prepared for the topography or shore survey and for the hydrography or survey of the water area.

Topography. The topographic survey of the shore and as much of the adjacent area as is required is usually made with a plane table, on which the map is actually drawn in the field as the work progresses. Points are located on the plane table sheet either by direct reading of the distance on a stadia rod or by intersections from two or more stations. On the plane table sheet it is customary to locate the shore or high-water line, the low-water line, off-lying rocks, streams, rivers, roads, towns, lighthouses, and all prominent features near the coast. Elevations are measured with the plane table or obtained from the triangulation, and are represented on the sheet both by figures and by contours, which are lines joining together points of the same elevation. For instance, a 100-foot contour represents the line where a plane 100 feet above sea level would cut the surface of the ground. It is particularly important in this topographic work to locate accurately objects which are good landmarks and likely to be of use to the mariner. In some regions auxiliary methods are used in filling in the topography, as, for instance, along a difficult coast each feature of importance may be located by sextant angles, or a traverse line may be run along the shore by the transit and stadia method.

The hydrography, or the survey of the water area, is of prime importance for the chart, but in the order of prosecution of the work it is convenient but not essential that it come after sufficient points have been located by the triangulation and topography. A hydrographic sheet is prepared on which all the points are plotted which will be useful. A system of sounding lines is then run over the entire area to be surveyed, locating the position of the sounding boat at intervals by sextant angles on survey signals or by angles from the shore. The ordinary method of sounding is to cast a lead from a boat and read the depth when the lead touches bottom and the line is vertical, and make note of the nature of the bottom. There is a systematic spacing between the casts of the lead and between the lines passed over by the boat, depending on the depth of water and character of the bottom. For soundings in deeper water various forms of sounding machines are used, with weight attached to a wire. For very great depths a small steel wire is employed and the weight is detached and left on the bottom. The deepest sounding thus far made, 5269 fathoms, or nearly six miles, was obtained by this method in the Pacific Ocean near Guam.

The offshore soundings are made from a surveying steamer; the inshore work is usually done by a launch or small boat.

So far as the navigational use of charts is concerned it is important that the hydrography shall show the limiting depths and the freedom from dangers, of channels, entrances, harbors, and anchorages. It is also desirable that the soundings shall be carried off shore at least as far as the one-hundred-fathom curve, as with the modern forms of navigational sounding machines it is possible for vessels under way to obtain soundings to this depth, and such soundings may be of value in identifying the location of the vessel. For depths greater than one hundred fathoms the soundings have less direct value to navigation except as proving the absence of shoaler areas, but soundings throughout the oceanic regions are of great geographical interest as well as of direct practical value in the laying of cables.

It is obvious that the plan of mapping the sea bottom by dropping a lead at intervals over its hidden surface is far from an ideal one. The lead gives the depth only at the point at which it touches the bottom, and no information as to the space between the casts except such as may be inferred from the relation of successive soundings. In numerous cases, after what was considered a very thorough survey of a region had been made, at some later day a pinnacle rock or other danger has been discovered. For instance, a very detailed hydrographic survey of Buzzards Bay was made in 1895; the sounding lines were run at intervals of 50 to 100 yards, and 91,000 soundings were made for a single sheet. Within this area the cruiser Brooklyn in 1902 touched a rock which was found to have 18 feet over it. (Fig. 17.) The least depth in the vicinity developed in the original survey was 31 feet.

For the satisfactory development of hydrographic work some invention is much needed which as it passes along the bottom will give a continuous depth curve. Several devices have successfully accomplished this in shoal water, but great credit awaits the inventor who designs something of more general application.

Tides and currents. Information must be obtained as to the movement of the water, both vertical and horizontal. The rise and fall of the tide are obtained by tide gauges, either automatic, which draw a continuous tidal curve on a roll of paper, or simple tide staffs, which must be read at intervals. The currents, whether due to the tides or other movements, are measured by noting the movement of partially submerged floats. Less accurate but useful information as to currents is obtained from the logs of vessels.

Dragging for dangers has long been resorted to for the investigation of isolated spots. A valuable and successful means has been employed recently of making sure that an area is free from shoals or rocks having less than a certain depth. This is done by dragging through the water a wire from 500 to 1400 feet long, and suspended at the required depth, with suitable buoys and weights, and kept taut by the angle of pull. If, for instance, the wire is set at a depth of 30 feet it will indicate the presence of any obstruction of less depth by catching on it and upsetting the buoys, and such spots are at once marked and investigated. Considerable work has been done with such drags in the last few years on the Atlantic and Gulf coasts and on the Great Lakes. This is of course a somewhat tedious process and gives no information as to depths greater than that for which the wire is set, but the experience already had indicates its great value. It will probably be found desirable in time to thus drag all water areas important to navigation where the depth is near the draft of vessels and the irregular nature of the bottom gives indication of dangers. In extensive dragging operations near Key West and in Jericho Bay, Maine, a number of shoals have been picked up which were not found in the original surveys.

A remarkable instance of the value of the drag was the recent discovery of a rock in Blue Hill Bay on the coast of Maine. This rock has but 7 feet of water over it, and is only 6 feet in diameter at the top. It is surrounded by depths of 78 feet, from which it rises nearly perpendicularly. The original survey gave no indication of a danger here, and its existence was not suspected until it was discovered with the wire drag.

Another method of dragging that has been employed is by means of a pipe suspended beneath a ship's bottom.

Magnetic variation. As the compass is a universal navigational instrument, information as to the magnetic variation is needed for the charts. The angle between the direction of the magnetic needle and the true north is measured at various points on both land and sea, and at some stations these observations are repeated after a number of years. From these results magnetic maps are made, from which both the variation and its annual change may be taken.

Reports of dangers. Aside from the more systematic surveys as outlined above, much information has been placed on the charts from other sources. On the earlier charts and on those of more remote regions at the present day much of the work has been sketched rather than surveyed. Even in the better surveyed portions reports come in as to dangers or other matters not shown, and if of importance and the report appears to be reliable these are sometimes at once put on the chart pending further investigation, or in other cases an examination is first made.

Shoals, rocks, and even islands have in numerous instances been shown on the charts which no one has been able to find again, and many of them after repeated searches have been removed. The same island or danger has sometimes been charted in two or more different positions as reported at various times. The treatment of such cases is one of the serious and interesting problems of the chart maker. It is generally less harmful to show a danger which does not exist than to omit one which does exist. On the other hand a non-existing danger shown on a chart may be the cause of actual expense and loss of time in compelling a vessel needlessly to go out of its course.

It is surprising to note with what lack of care and of sufficient evidence reports of dangers at sea have sometimes been made, and how incomplete are many of the reports even when the existence of the danger is beyond question. It is unfortunately true that some of these reports are the result of effort to escape blame for accident by throwing the fault on the chart. Many such reports also result from various illusory appearances. A large tree covered with weeds, an overturned iceberg strewn with earth and stones, a floating ice-pan covered with earth, the swollen carcass of a dead whale, a whale with clinging barnacles and seaweed, reflections from the clouds, marine animalculÆ, vegetable growth, scum, floating volcanic matter, and partially submerged wrecks covered with barnacles, have been mistaken for islands, shoals, or reefs. A school of jumping fish has given the appearance of breakers or caused a sound like surf, and tide rips have been mistaken for breakers. Raper very properly calls attention to the obligation upon every seaman of carefully investigating doubtful cases and making reliable reports. "Of the dangers to which navigation is exposed none is more formidable than a reef or a shoal in the open sea; not only from the almost certain fate of the ship and her crew that have the misfortune to strike upon it, but also from the anxiety with which the navigation of all vessels, within even a long distance, must be conducted, on account of the uncertainty to which their own reckonings are ever open. No commander of a vessel, therefore, who might meet unexpectedly with any such danger, could be excused, except by urgent circumstances, from taking the necessary steps both for ascertaining its true position, and for giving a description as complete as a prudent regard to his own safety allowed."

As to the older doubtful dangers now shown on the oceanic charts, it is estimated that the positions may be considered as uncertain by 10 miles in latitude and 30 miles in longitude, and areas of this extent must be searched to determine definitely the question of their existence.

The following are interesting or typical cases of reported dangers:

The master of an Italian bark in September, 1874, reported sighting a large rock in latitude 40° N. and longitude 62° 18´ W. Fortunately for the charts there were two independent reports from other vessels in the same month of sighting a partially submerged wreck in this vicinity.

The Spanish steamer Carmen was wrecked in 1891 by running on a rock off the southwest coast of Leyte; the rock was reported to lie one mile off shore, a dangerous position for vessels using Canigao Channel. A survey made in 1903 showed 58 feet of water in this location, and that Carmen Rock on which the vessel struck was really within one-fourth mile of the beach. The rock had, however, for twelve years been shown on the charts in a position which made it an obstruction to navigation.

The ship Minerva in 1834 was reported to have struck a rock near the middle of the broad entrance to Balayan Bay; the fact that this occurred at 2 A.M. indicated a very doubtful position, but it was stated that an American ship had previously been wrecked on the same rock. It consequently appeared as a danger on the charts for seventy-one years, when a survey showed no depth of less than 190 fathoms in this vicinity, and it was removed from the charts.

A British steamer was wrecked in San Bernardino Strait in 1905; the master reported that he was in a position where the chart showed 51 fathoms, and that he was 112 miles distant from Calantas Rock, and on these grounds the finding of the official inquiry was that "no blame can be attached to the master, officers, or any of the crew for the casualty." Very shortly after the disaster, the surveying steamer Pathfinder definitely located the wreck and made a survey of the vicinity. The previous chart of Calantas Reef was found to be fairly correct, and the stranding was determined to have occurred well within this reef in a position where the chart showed soundings of 334 to 434 fathoms, and 12 mile from Calantas Rock, which rises 5 feet above high water.

A transport entering San Bernardino Strait a few years ago ran on a rock and was damaged; the position was reported as about two miles southeast of San Bernardino Island and near the middle of the passage. The rock was not put on the charts, as prompt investigation showed 50 fathoms of water in this vicinity, and that in all probability the transport actually touched a small reef making out from the island.

The master of the brig Helen reported that his vessel was wrecked on a reef lying six miles from Rockall. When surveyed Helen Reef was found to be about one-third this distance from Rockall.

An island has been reported in eight different positions, ranging in latitude from 30° 29´ to 30° 42´ N. and in longitude from 139° 37´ to 140° 38´ E.

There have been a number of reports of islands in the area from latitude 40° 00´ to 40° 30´ N. and longitude 150° 30´ to 151° 00´ W. The master of the bark Washington reported in 1867: "On my passage from the Sandwich Islands to the northwest coast of the United States, when in latitude 40° 00´ N., in a dense fog, I perceived the sea to be discolored. Soundings at first gave great depths, but diminished gradually to 9 fathoms, when through the mist an island was seen, along which I sailed 40 miles. It was covered with birds, and the sea swarmed with seal and sea elephants." A United States vessel searched in this vicinity without seeing any indication of land, and obtained soundings of 2600 fathoms. A British ship in 1858 searched for fourteen days over this area without finding anything. Searches were also made in 1860 and 1867 without success, and the present charts show no islands in this part of the Pacific.

In a number of cases erroneous positions have been due simply to blunders. Thus Lots Wife, first seen by Captain Meares in 1788, was shown on his chart in latitude 29° 50´ N., longitude 156° 00´ E., and stated in his book to be in latitude 29° 50´ N. and longitude 142° 23´ E. Massachusetts Island by one report was in longitude 177° 05´ E. and by another in 167° 05´ E. The apparent blunder of 10° is now immaterial, as the island has disappeared from the charts altogether. The Knox Islands were placed by the Wilkes Exploring Expedition in latitude 5° 59´ 15´´ N., longitude 172° 02´ 33´´ E. The old British charts showed islands of this name also in latitude 5° 59´ N., longitude 172° 03´ W., the longitude being doubtless transposed. In the case of Starbuck Island, discovered south of the equator, the latitude was apparently transposed, as on old charts it was also shown in the position, latitude 5° 40´ N., longitude 156° 55´ W.

A pinnacle rock can sometimes be located only with great difficulty even when known to exist. Rodger Rock, on which the bark Ellen struck and was damaged, lies in latitude 0° 41´ 15´´ N. and longitude 107° 31´ E. It has but three feet over it at low tide. The British surveying ship Rifleman searched four days before finding it, although the plotted tracks showed that she and her boats had passed very close to it. This indicates that great caution must be used in removing a reported danger from the charts.

The old charts of the Atlantic indicated a danger 30 to 45 miles to the southwest of Cape St. Vincent. This danger was omitted from the charts about 1786 owing to lack of confirmation. Later, in 1813 and 1821, it was reported that vessels were lost or damaged by striking this rock. Soundings of over a thousand fathoms are now shown on the chart in this vicinity and the rock no longer appears.

A comparison of a Pacific Ocean chart of about forty years ago with one of the present time (Fig. 19) illustrates in a striking manner how many doubtful dangers, or vigias, have gotten on the charts and how after laborious search many of them have now been removed. This condition was especially true of the Pacific, owing to the numerous reports of an indefinite nature from whaling ships, among whose captains there was a saying "that they do not care where their ship is, so long as there are plenty of whales in sight."