12 



Let us finally consider oscillations. Roll experiments are a well- 

 established requisite of the testing technique. However, the author does not 

 know of any systematic experiment dealing with k as a function of the speed 

 of advance cr the frequency of roll. It has been reported without further 

 explanations that a variability of the period with the Proude number has been 

 observed on a model of the Conte di Savoia. 11 Probably it would be worth- 

 while to check the few pertinent publications if a dependence of k upon F 

 can be stated; obviously, possible variations of the metacentric height must 

 be known. 



Some twenty years ago I recorded extinction curves of the heave and 

 pitch motion; the work was carried out at the Berlin Model Easin. 12 We in- 

 vestigated a full and a moderately-full model. V.'ithin the range of Proude 

 numbers < F < 0.20 one cculd not find a dependency of the inertia coeffi- 

 cients k , k and of the damping 26 , 2 6, upon the speed. The measured 



z ~ yy~ z yy 



values of k , k were slightly less than those computed by the strip method 

 with ellipsoidal corrections" (see Section 1.4). Later, similar experimental 

 results for ship models have been published only occasionally. 



More detailed heaving experiments with geometrically defined bodies 

 have been performed by Dimpker 13 and Holstein. 14 Although Holstein 1 s experi- 

 ments were carefully conducted, their accuracy is not high enough to establish 

 a consistent dependency of the added mass values upon P w . Following Wendel, 

 the experimental values of k plotted over the beam-draft ratio are somewhat 

 lower than the corresponding "theoretical" values of k . Holstein was unable 

 to detect the hump in the range of small P w since the frequencies tested cor- 

 respond to a range F w > 1 .5. 



As has been mentioned before, the usefulness of the added mass con- 

 cept can be questioned when its value becomes a function of many variables. 

 However, it can be asserted that the concept remains important since 



1 . With increasing depth of immersion f the values k approach quickly 

 the- "classical" values k. 



2. The order of magnitude of k is generally identical with that of k. 



3. The solutions k, k, and k obtained under simplified conditions, 

 yield over large ranges, valuable estimates for k. 



We mention two special problems, shallow water effects and gravity 

 effects. In a well-known paper I.I. Koch has calculated the inertia factors 

 k and k . of a rectangle, using an electrical analogy. 15 Schnadel has suc- 

 cessfully applied these results to the interpretation of vibration experi- 

 ments. 16 Since gravity effects so far have not been considered, we restrict 



