NO. 15.] EXPERIMENTAL ERRORS AND CORRECTIONS. 87 



D. EXPERIMENTAL ERRORS AND CORRECTIONS. 



On account of the small scale on which the experiments had to be carried 

 out, even a small error in the determination of resistance might be of im- 

 portance in applying the results to the case of full-sized ships. Most of these 

 errors have, however, practically the same value (at a given velocity of the 

 boat) independently of whether the boat moves with a small resistance in 

 homogenous water, or with a great resistance in "dead-water". They have 

 consequently no essential influence upon the determination of the "dead-water 

 resistance" (see p. 61); and this is the part of the resistance which is of the 

 greatest interest to us. An approximate determination of the resistance in 

 homogenous water is sufficient for our purpose. 



With regard to the measurements in homogenous water, as well as in fresh layers 

 above salt water, the following points may be noticed. 



1) The measured resistances are somewhat too high, owing to the friction against the 

 steering string (p. 59). To approximately determine it, some experiments in homogenous 

 water, were made without, as well as with, steering string (see experiments 18—47, p. 78). 

 The friction, found from these experiments, is represented graphically by the faint curve 

 in Fig. 1, PI. X. As it was difficult to make the boat go straight, without using the 

 steering string, the determination of the friction is only approximate. It was eventually 

 found most convenient not to make any correction for it, and the values of resistance given 

 in the tables and on the diagrams, therefore include friction. Exception must be made in the 

 case of most experiments with the small .Fram-models, in which the steering string was 

 not used. When steering string was not used, it is always mentioned in the tables. 



2) Tlie friction in the wheels of the towing-apparatus, was as already mentioned, al- 

 lowed for when adjusting the towing-weights. This friction varying from 16 mgr. in the 

 case of the smallest towing-forces, to 82 mgr. with a towing force of 10 gr , was by 

 control-determinations found to be practically the same at the end of the experiments in 

 June 1901 as in November 1900. It is consequently without influence upon the measurements. 



3) The loss of towing-force owing to the small ^inclination of the towing-string from 

 the horizontal line, was, when calculated, found to be quite insignificant. 



4) The dust which was constantly settling on the surface, was a source of resistance. 

 It gathered up ahead of the boat and] formed a rigid disc, which was pushed before 

 just like an ice-floe would be. The resistance which it caused, seemed to be independent 

 of the velocity of the boat, but it increased with the quantity of dust on the surface. The 

 "dust-floe" grew larger and thicker, as the boat moved from one end of the tank to the 

 other; and the resistance increased correspondingly. If there was very much dust, it would 

 even stop the boat in spite of a towing force of half a gramme or more. To get rid of this 

 error, which was noticed in the very first experiments, the dust was skimmed off in the 

 manner described on p. 57. This operation ] was made generally twice a day, sometimes 

 more often; and in this way, the errors due to dust, were kept within 2 or 3 cgr., in 

 any measurement of resistance inserted in the tables. In most cases, it has probably not 

 been more than a centigramme. 



5) If the surface be even very little contaminated, so that its surface-tension be reduced, 

 it will cause an increased resistance, as was mentioned on p. 53. In the eddying wake 



