50 APPENDIX. 



DEDUCTIONS FROM THE PRECEDING EXPERIMENTS. 



\ 



In experiments such as the preceding, made on a small scale, there frequently arises a 

 difference of action, and consequent effect, at variance with the true practical results : it 

 therefore becomes important, before drawing conclusions from such experiments, to examine 

 into the causes, which will operate differently on the large and small scale. In the model of 

 Morgan's wheel the friction will be proportionally greater than in the actual wheel : the 

 model was however made so perfect, that it could not have been much greater, although 

 some allowance should be made. 



The cycloidal wheels lost much of their effect from being on the small scale : the intervals 

 between the paddles were so small, that the water could not escape, owing to its natural 

 adhesion, which would have no sensible effect in practice : the loss from this was evidently 

 so great, as could be seen by examining the motion of the wheel, that no just comparison can 

 be made of its effect as compared with the others. These considerations will not, however, 

 affect the comparison of the same wheel in different states of immersion, from which we learn 

 the nature of its action ; nor will it affect the comparison with the common wheel with different 

 number and surface of paddles ; so that many conclusions of practical utility may be derived 

 by an examination of the preceding table of experiments. 



The most direct comparison of the action of Morgan's with the radiating wheel, is with the 

 last experiment, No. 10., which is in fact the same wheel with the paddles fixed radiating. 

 Here, in the small immersions, the effect of Morgan's wheel is less than the radiating, the 

 mean of the three first immersions of the former being 6990, and of the latter 7156 ; which is 

 owing to loss, as we have before pointed out, from the additional receding of the wheel : this 

 is rendered strikingly evident from the velocity (in the experiments) on Morgan's wheel being 

 at the same time greater than in the other. 



In the deep immersion, Morgan's wheel has considerably the advantage, not only in velo- 

 city, but in effect, the mean of the three last observations on the former being 6790, and the 

 latter 5003 ; the loss from oblique action operating in this case very much against the com- 

 mon wheel. 



In examining the other experiments, we find that a less deep paddle with a greater number 

 of boards gives a better effect throughout. In comparing the best radiating wheel with Mor- 

 gan's, we find it is completely superior in the small immersions, and continues to be so until a 

 greater proportion than '235 of the diameter of the wheel is immersed. In the experiments 

 an allowance should be made for the loss of effect from the greater friction of the model ; at 

 the same time, there is little doubt that a well-proportioned common wheel will be superior to 

 Morgan's in small immersions. 



The experiments on the cycloidal wheels were not so satisfactory as could have been 

 wished, from the loss of effect evidently produced by the adhesion of the water in coming 

 out, a difficulty that could not exist on the large scale : they however show a greater 

 equality of effect throughout the different states of immersion than any of the wheels ; arid, in 

 fact, in the deep immersions are superior to most of the wheels ; which ar,e important points 

 in favour of this construction, and leave little doubt that the deficiency in the light dips, 

 from the cause above named, was very great. 



