the Wild T-2 theodolite, and the measurement was made with three Lovar tapes. Thermometers and stretcher apparatus of an approved type were used, The calibration certificates of these tapes are included in the record of the survey. Stakes were set at fifty meter intervals for chaining points, and the tapes were alternated so that in completing the forward and backward measurement all three tapes were used in each direction, Because of velocity of the wind at this period of the year, it was necessary to use a windbreak in order to obtain accurate results. This consisted of a thirty-six inch wide strip of canvas approximately sixty yards long, held parallel to the line as each measurement was made. Highteen men were required to hold the windbreak in place, The manpower ai the time required to measure this base line would have been greatly re- duced had it been possible to make the surv2y during the more favorable weather of the summer months. The computed probable error of the total measurement is 1 part in 648,000. The allowable maximum probable error is one part in 500,000. VERTICAL CONTROL Because no records were available of vertical control established by previous surveys, a temporary datum was established on each of the Project islands as surveys were made. For each island, a datum approximating mean low water springs was arrived at by applying corrections from the U., S. Coast and Geodetic publication "Tide Tables of the Pacific Ocean" to a series of tidal observations. This gave a temporary datum, significant to less than a foot. The procedure was to erect a tide staff at a suitable location at each island to take periodic observations as surveys were made at these islands. After corrections had beon applied, a mean of these corrected observations was accepted as the temporary datum. This datum was trans- ferred to a permanent momment in the vicinity and became the origin of all vertical control on the particular island. Although the vertical datum of one island bore no exact relationship to that of another, there was no requirement that made knowledge of such relationship essential. Such inaccuracy as existed was of no con sequence, inasmuch as most any one island is not related to any other. Caution was taken that no tidal observations for establishing verti- cal datum were taken at times of storm, unusual high winds, or surge. It was observed that at such times high tides would be between six and seven feet in elevation, whereas the predicted tide might be four and one-half feet. No continous study of tides was made. As stated, the datums established were sufficiently accurate for Project requirements. However, it is recommended that as future work is carried on, these datums be refined and the inter-relation of the datums 5-66