Observers' Field Reports 99 



railway point of the itinerary. The following day I started for Goyaz City with pack 

 animals. The 320 miles from Catalao were covered by May 4. 



After an examination of all available information regarding routes, it seemed best 

 to proceed first by pack animals to Leopoldina on the Araguaya River, thence down this 

 river to Conceicao, a town not shown on the maps, thence westward overland to the Rio 

 Fresco, an uncharted branch of the Xingu, and finally to follow these waters to the 

 Amazon. The journey started May 18, and Leopoldina, on the Araguaya, was reached by 

 pack animals May 25. 



Finding that the Araguaya could be ascended readily as far as the little town of 

 Registro do Araguaya, I decided to make this trip in the hope of finding some route to the 

 upper Xingu, which would be practicable with the time and means at my disposal, as my 

 instructions particularly mentioned this route. Accordingly, I started up the river in a 

 small boat, with a crew of 3 men, and arrived at Registro June 8. As no satisfactory in- 

 formation could be obtained regarding routes to the headwaters of the Xingu, we re- 

 turned to Leopoldina and arrived June 12. On June 14 the descent was resumed, and 

 Conceicao was reached July 8. Here it was found that a route to the Rio Fresco and 

 Xingu was feasible, although expensive. However, I considered it inadvisable under the 

 circumstances, and, therefore, accepted an offer to accompany a rubber trader who was on 

 the point of leaving for Para (Belem) by way of the Araguaya in a well-manned barge. 



On July 11 the descent of the Araguaya was continued. The boatmen proved ex- 

 pert, as the first of a numerous series of rapids was encountered on July 14. From July 14 

 to August 18 we had many interesting experiences, grounding in shoal places, shooting 

 numerous rapids of all sizes and degrees of difficulty and danger, and making portages. 

 (See view 4 of Plate 6.) 



On August 18 we arrived at Alcobaca where there is steamer communication with 

 Para. Accordingly, travel was resumed by river steamer on August 22, and Para was 

 reached August 30. With the reoccupation of our magnetic station at Pinheiro, near 

 Para, the work was concluded. On September 4, I sailed for New York, arriving at 

 Washington September 22, 1915. 



The time consumed in carrying out this work was 200 days (March 6 to September 

 22, 1915). Excluding travel to and from Washington, 169 days were spent in the field, 

 so that an average of 4.1 days was required for a station, but if the intercomparisons at 

 Vassouras are omitted and field time counted from the first campaign station at 

 Araguary, then the average time per station is 3.7 days. The only delay experienced was 

 due to the numerous rapids in the lower course of the Araguaya and Tocantins. The 

 Itaboca series in the Tocantins alone consumed ten days. Travel to and from the field 

 amounted to about 8,050 miles, travel in the field to about 2,550 miles, of which 725 were 

 by rail, 450 by pack animals and about 1,375 by river. About 200 miles of the river 

 travel were covered by steamboat. The average total travel per station was 258.5 miles, 

 and the average field travel was 62.2 miles per station. 



It is a great pleasure to make acknowledgment here of the courtesies and assistance 

 extended by Dr. Henrique Morize, director of the National Observatory of Brazil; by 

 Mr. Gottschalk, the American consul at Rio de Janeiro; by Mr. Horace E. Williams of the 

 Brazilian Geologic Service; by Mr. Frederick Glass, English missionary at Goyaz City; 

 and by Frei Sebastiao M. Thomas of Uberaba, Inspector of Dominican Missions. 



The course of the Araguaya between Registro and Leopoldina is represented by 

 dotted fine on the Mappa Geral do Brazil of 1913, and Stieler's Atlas. The whole part 

 south of the Rio das Mortes appeared as Rio Araguaya or Grande on Stanford's map, 

 which contained less detailed information than either of the other two. Since many of 

 the latitudes from these maps differ from our observed latitudes, it is probable that the 

 map-longitudes are also uncertain. 



