in the row, preceded by S) , and finally we state (in parentheses) 

 the total concentration, in terms of atmospheres of osmotic pres- 

 sure representing the calculated osmotic value of the solution in 

 question. Examples are:— IR2S3 (1.00 atm.), IIIR1S2 (1.50 

 atm.), VR6S1 (1.65 atm.), etc. This form of notation is em- 

 ployed in the tables given below, and it is hoped that all coopera- 

 tors will adhere strictly, in order to facilitate comparisons. 

 (Previous writers have employed C in placer of S, but it seems a 

 little better to use S as standing for solution rather than C, de- 

 noting culture.) 



Blank triangular diagrams, printed on sheets SV^xll in., may 

 be obtained from Baltimore at a price of $1.00 per hundred. 



Aside from the kinds of salt entering into one of these solu- 

 tions (solution type) there are four characteristics by which 

 that solution may be distinguished: (1) volume-molecular salt 

 proportions, (2) osmotic salt proportions (see Tottingham, 

 1914), (3) total volume-molecular concentration (proportional 

 to the number of molecules, of all kinds, per liter) , and (4) total 

 osmotic concentration (taken to be proportional to the number 

 of particles, — molecular groups, molecules and ions, of all kinds, 

 — per litre; this last characteristic may be expressed in various 

 ways, but the atmosphere will be the unit here used as a measure 

 of the calculated osmotic value, or the calculated potential 

 osmotic pressure (of which the solution is taken to be capable) . 

 On account of ionization (and probably hydration), nos. 1 and 2 

 do not vary proportionally to each other from solution to solu- 

 tion in our series ; nor do nos. 3 and 4. It follows that if we plan 

 our series of twenty-one solutions to vary by definite increments 

 on the basis of no. 1 (volume-molecular salt proportions), then 

 they must vary irregularly in respect to no. 2 (osmotic salt pro- 

 portions). The variation cannot be regular by both criteria at 

 once, and we have chosen the former as the one to be used. This 

 criterion (no. 1) can be definitely stated for any of the solutions 

 without any assumption regarding ionization, osmotic phe- 

 nomena, etc. On the other hand, since the total concentration 

 of a solution must act upon plants primarily in an osmotic way, 

 the calculated osmotic value (no. 3, above) is here used for this 

 measure of the main physical character of any solution. This 

 usage involves calculation, and assumptions as to ionization, or 

 direct measurements of the lowering of the freezing-point, but 

 it seems desirable (at the start, and in spite of uncertainties) 

 to compare all the twenty-one selected sets of salt proportions, 

 keeping the calculated total osmotic value constant throughout 



27 



