underestimate of the extreme wave-induced hull- 

 girder stress would "inevitably doom many ships 

 to the bottom" is believed to be rather on the 

 pessimistic side. There are many other stresses 

 present besides the wave-induced stress and not 

 all will be likely to experience their maximum 

 value at the same time. 



The comments of Mr. Darbyshire are par- 

 ticularly welcome as he has pioneered in the 

 measurement of ocean waves with a shipborne 

 wave recorder. His data provide the first ex- 

 tensive series of measurements of ocean-wave 

 heights. It is very gratifying to note that Mr. 

 Darbyshire finds^' that the proposed log-normal 

 long-term distribution fits the ocean data satis- 

 factorily. It might be well to note at this point 

 that the author did not study the short-term dis- 

 tribution of wave heights and made no attempt 

 to show that their distribution was of the Ray- 

 leigh type; rather he stated that "it has been 

 accepted generally that the heights of ocean waves 

 passing any given point follow the Rayleigh dis- 

 tribution. . ." Recent studies indicate that of all 

 the variables discussed here, the short-term pat- 

 tern of the height of the wave surface is least well 

 represented by the Rayleigh distribution. 



Both Mr. Comstock and Mr. Vasta express the 

 opinion that endurance strength is not a problem 

 in ship structures, inferring that only the maxi- 

 mum total stress need be considered. It is not 

 believed that there is a sound basis for this con- 

 clusion. A study by Dr. Week'' has shown that 

 fatigue failures of typical welded plate-stiffener 

 combinations may occur at very low nominal 

 stresses (±2 tons/sq in.) at the point of failure. 

 Such cracks, when noted, probably would be re- 

 paired without coming to the attention of de- 

 signers; they could on occasion form the nucleus 

 of a major fracture in the presence of large tem- 

 perature-induced, dead-load, or seaway stresses. 



The discussions of Week's paper force one to 

 one of two conclusions, quoting Dr. Week: 

 "Either that fatigue failures do not occur in ships, 

 which seems to be contradicted by the majority 

 of those who have contributed to the discussion, 

 or that they are not recognized as such. . ." It 

 also may be timely to quote the following from his 

 paper: "It is difficult to see why it should be 

 profitable to distinguish between the main hull 

 structure of a ship and the rest when considering 

 the incidence of fatigue. This distinction may 

 well be convenient for purposes of design, but it 

 does not correspond to reality." 



"A study of ship failures sometimes referred to 



as providing a basis for concluding that fatigue 

 failures are not important is the "Final Report 

 of a Board of Investigation" dated 15 July 1946, 

 which was made for the Secretary of the Navy to 

 study the high incidence of failures in welded ships. 

 The service data given in that report are heavily 

 weighted in favor of failures incident to the brittle- 

 fracture problem because the ships were es- 

 pecially subject to this "disease" and thus could be 

 expected to "die" thereof before an endurance 

 failure could develop. What is needed is a con- 

 sistent and thorough system of reporting struc- 

 tural failures and "defects" even though they may 

 not appear dangerous. The author did not in- 

 tend to convey the impression that "fatigue" is a 

 major structural problem in ship design — he does 

 not know — however, this problem should be 

 studied together with the other possibilities of 

 causing failure, without prejudice. 



The author is not in a position to answer Mr. 

 Vasta's question as to whether hull-girder struc- 

 tures are designed too conservatively as he has no 

 sound basis for a definite conclusion. However 

 the author "feels" that the sectional moment of 

 inertia is probably more than adequate in many 

 ships. On the other hand the constructional de- 

 tails with which the structural material is put 

 together may not always be adequate and further- 

 more the local strength, such as that of bottom 

 structures subjected to slamming, is on occasion 

 insufficient. It is really not realistic to separate 

 local strength from huU-girder strength. The 

 author fully appreciates Mr. Vasta's other com- 

 ments and is in general agreement with them. 



Mr. Marks is increasing the scope of the paper 

 when he states that it attempts to solve the ship- 

 performance problem. This is not so. The ob- 

 ject, as has been stated repeatedly, is to establish 

 that the long and short-term distribution of ship 

 motion and stresses approximate to the Rayleigh 

 and log-normal distributions, respectively, and to 

 give illustrative engineering applications of this 

 result. True, it was hoped that a knowledge of 

 the form of the distribution would simplify the 

 solution of the ship-performance problem. As 

 stated in the summary, it is expected that this so- 

 lution will take the form of model tests in waves 

 which will be utilized to make predictions for 'par- 

 ticular ships in a particular service analogous to 

 the procedure now used to determine power re- 

 quirements in smooth water by model tests. 



Professor Korvin-Kroukovsky feels that in the 

 paper there is a lack of reference to the physical 

 nature of the situation. The physical nature is of 

 course reflected in the parameter E (the mean 

 square value of the physical motion) which re- 

 lates the mathematical distribution to the "physi- 



56 



