Geographic Positions 25 



of three or more high-grade watches; (d) observations of occultations or of eclipses of the 

 moons of Jupiter; and (e) utilization of data from reliable maps and previous surveys. 



At all stations, unless prevented by lack of time or by cloudy weather, observations 

 on the Sun, or on a star, are made to obtain the correction of the time-piece on local mean 

 time. For most of the work done, it is not possible to exercise time-control by method (a) . 

 The question of a suitable radio-outfit (method b) has received consideration with respect 

 to the special needs of the more recent expeditions. On a few expeditions a box chro- 

 nometer has been carried, but with unsatisfactory results, owing to the difficulties of the 

 methods of transportation available. Night work, such as required by (d), is usually 

 objectionable, especially in the tropical regions, where much of the Department's work 

 has been done, on account of risk to the observer's health and to the success of the ex- 

 pedition. Since, furthermore, the observations by (d) are long and troublesome to reduce, 

 and can only be made at predicted times, without opportunity for desired repetitions and 

 checks, no very serious attempts have been made to use occultations, or similar astro- 

 nomical methods, for the determination of longitude. Some regions are so well mapped 

 that the required longitudes may be scaled from the maps with sufficient accuracy; thus, 

 for the extensive work in Australia, satisfactory geographic positions could be obtained 

 with the aid of the excellent system of surveys covering most of that country. 



After careful study of the conditions and of the experiences gained on numerous ex- 

 peditions, transporting time by means of three or more watches, has been the method 

 generally adopted and in many cases has given very good results. The best of watches, 

 however, for one reason or another, often become unreliable when subjected to the trying 

 conditions of a field expedition extending over several months. In such cases, the longi- 

 tudes of the most important points as obtained from the best available sources are ac- 

 cepted, and the intermediate positions are derived, with the aid of the determined watch- 

 rates, by interpolation. 



In order to keep the watches at a fairly constant temperature, the observer during the 

 day usually carries them in a belt on his person. He keeps them upright during the night , 

 winding them daily at about the same time, and comparing them with each other morning 

 and evening and whenever necessary for control. With not less than three good time- 

 pieces thus treated, consistent longitudes have been derived in regions where no good 

 values were otherwise available. The experiences, for various reasons, have been un- 

 satisfactory when the attempt was made to use a least-square formula for reducing the 

 results from the individual time-pieces. 



It is recognized that all time-pieces of a set may be affected alike when subjected to 

 identical treatment, e. g., to the same changes in temperature and to the same vicissitudes 

 of transportation. They may all run faster or all run slower, and no method of determin- 

 ing a rate from comparisons between them will serve to detect that fact. Using watches 

 of different sizes, different makes, and varying adjustments reduces the error apppreciably 

 but does not eliminate it. 



In general, a careful analysis of the performance of the time-pieces as shown by the 

 daily comparisons and the observations for local time will, for short expeditions, so con- 

 trol obvious changes in mean rate and occasional abrupt changes that watch-corrections 

 on standard time may be derived which will serve the purpose for which the observations 

 are made, though falling short of the accuracy desired by the geographer. For such ex- 

 peditions the maximum error is often apparently kept within 5 to 8 seconds, though it 

 will doubtless largely exceed this under unfavorable conditions. It may be of .interest 

 to note that the best maps available for many of the regions traversed by our observers 

 have been found in error in remote places by as much as one-half degree, correspond- 

 ing to 2 minutes of time. 



