wood models. These are glued together from seven-eighths-inch white pine lifts cut to shape on the band saw and finished by hand to transverse templates.
Improvements in technique and apparatus have been made as required, but when funds are available a more modern dynamometer and car will be installed. The University of Michigan Tank has been in great demand in the development of barges and towboats to operate on the Mississippi and Ohio rivers because of its 140-foot false bottom, which can be adjusted in depth to simulate shoal water conditions. The Tank has been used in a study of the correction of yaw, a condition prevalent in barge towing, and much attention has been given to the improvement in speed and fuel economy of existing lake bulk carriers. This work has been under the direction of Baier.
Sensing the need for closer co-operation and standardization of technique among the various tanks, Baier organized in 1938 the American Towing Tank Conference (ATTC). The first meeting was held at the Stevens Institute of Technology, Hoboken, in April, 1938.
Although open-water research has been carried out on propellers and paddle wheels, self-propelled model tests have never been undertaken by the University. It is believed that this type of work should be restricted to models of at least twenty feet in length in order to avoid scale effect, and these sizes are beyond the present Tank's capacity.
Among the research problems undertaken have been the following. Series 1050 was tested in 1921 for the Fairbanks-Morse Company to find the best forms for fireboats. A fireboat is usually made as short as possible to enable it to maneuver readily in docks, and usually it has a high speed relative to its length. The forms were designed by Alfred J. C. Robertson, and the series consisted of four models all cut from the same waterlines. This series was unique in having the vertical spacing changed. The results obtained were very satisfactory throughout the speed-length ratio for which they were designed.
The first tests in the 1130 series were undertaken at the request of the United States Shipping Board in 1919. Nine preliminary forms were tested, and form 1130 was selected. Nineteen models were made from a set of parent lines and tested with varying percentages of entrance and run. This series was planned by Robertson.
At the University of Michigan four more models with a 50 per cent run were tested in order to complete the series. This was as far as it was originally intended to go with tests on this model, but the question arose concerning the effect of section shape upon resistance, and it seemed best to use model 1130 for such experiments. A medium shape was used on the twenty-three models previously tested, and variations were made in both directions. The tests seemed to indicate that a 27 per cent entrance and a 40½ per cent run were about the best proportions for general purposes.
All the tests had been made for a 425 by 56 foot ship. The next logical variation seemed to be in the beam. Kent's paper in 1919 on the "Effect of Variation in Beam" had covered a wide range, there being 13-foot intervals between the beams of the successive ships tried. It seemed advisable to cover a more limited range and to have about 4-foot intervals. Thirty models were used to cover this ground, and each model was tested at four different drafts. The experiments upon shape of section having shown that for this type of model best results were obtained with V-shaped forward sections and bulbous-shaped stern sections, these thirty models were tested with V-shaped sections forward and bulbous-shaped sections aft. Thirty-six models were used to
