ACTIVE TYPE OF STABIUZING GYRO. 219 



oscillations not caused by resonance are of a very small amount only. This has 

 been confirmed by rolling diagrams taken from tank experiments as well as from 

 a great number of ships. On Plate 17 of his above-mentioned paper, Herr Frahm 

 has published curves from tank experiments, showing the behavior of a ship when 

 struck by quite irregular impulses, their period varying from 5 to 1 1 . The diagrams 

 taken on board ships also show that forced oscillations of ships caused by waves 

 of a different period make the ship swing quite irregularly, but do not produce 

 heeling amplitudes comparable to those of natural rolling. Even on smaller 

 vessels these amplitudes scarcely amounted to more than 5 or 6 degrees. The 

 conclusion to be drawn is that any anti-rolling device has to be designed ONLY for 

 the case of resonance. I think there is a discrepancy in Mr. Sperry's paper con- 

 cerning the question of resonance. If he doubts the effectiveness of tanks in 

 "instances only slightly removed from synchronism," why then does he assume 

 heavy rolling of a ship if "there is no such thing as resonance between waves 

 and ship?" 



To prove his statements of the ineffectiveness of the tanks upon conditions 

 where the sea is out of synchronism with the period of the ship, Mr. Sperry takes 

 the curves on Plate 80 of his paper. I think these curves, showing the rolling 

 amplitude of the ship for a very few minutes only, extend over a time much too 

 short to give a fair view of the usefulness of the tanks. The original curves from 

 which Mr. Sperry has taken the curves shown on Plate 80 are published in Herr 

 Frahm's paper and extend over a period of several hours for each instance. More- 

 over, they are taken at different loading conditions of the two sister ships Ypiranga 

 and Corcovado, which, of course, affect the period of the ship. The periods indeed 

 vary from 3.73 to 4.61 oscillations per minute in the case of the Ypiranga, and 

 diagrams taken for different periods of the ship prove that the great usefulness 

 of the tanks is not considerably affected by the change in the period of the ship. 

 The reduction of the amplitudes is in one instance from 1 1 degrees to 2 1 to 2 degrees, 

 in another instance from 18 degrees to 4 to 3 degrees, and with only one tank in 

 action from 18 degrees to 7 degrees, which is much more than in the curves on 

 Plate 80 of Mr. Sperry's paper. Since some of Herr Frahm's diagrams have been 

 taken in a very heavy sea gale, I think Mr. Sperry's assertion that "the percentage 

 of the departure where they fail grows smaller and smaller as the sea grows heavier," 

 is not justified, and I am sorry Mr. Sperry did not publish any data or curves to 

 prove his words, "We find that the tanks have already practically failed to perform 

 any useful function whatever." Since the time of the publication of Herr Frahm's 

 paper, rolling diagrams have been taken on a great number of ships, all of them 

 showing a very large reduction in the heeling amplitudes, though the tanks have 

 never again been placed as high as in the case of the Corcovado and the Ypiranga. 

 In the Kaiserin Auguste Victoria of the Hamburg-American line, for instance, 

 the rolling amplitudes in a very heavy storm were reduced to 33 per cent of their 

 amount when the tanks were not in action. This is no specially selected example, 

 but one which just occurs to me. 



