10 



solution. In other words, in the case of monovalent compounds, one 

 gram-molecule was contained in a liter of solution, while in the case of 

 bivalent compounds, a half gram-molecule was present.^ In this way 

 only is a really instructive and fair comparison of the effects of different 

 salts obtainable. Many experiments made in times past in which com- 

 parisons were based upon simple percentages of solute to solvent by 

 weight are for this reason of far less value than if normal solutions 

 had been employed. In order to study comparatively^ those effects 

 produced bj^ different electrolj^tes which are not dependent ujjon 

 bheir respective chemical natures, but which are common to them all 

 and due onlj' to their active masses (such, for instance, as effects due 

 to the osmotic pressure existing in the solution), it is obviously neces- 

 sary to take into consideration the number of reacting weights of the 

 electrolj^te introduced and the amount of electrolytic dissociation 

 which takes place. That is to sa}^ one must consider the concentra- 

 tion of the solution with respect to the number of reacting chemical 

 equivalents, molecules, or ions which may be present. Moreover, 

 attempts to studj' comparatively the effects produced by different 

 kinds of ions in the solution can onlj' be made by approaching the 

 subject in this manner. But in all statements in this report of the 

 concentration of a given solution both fractions of a normal solution 

 and parts of salt to 100,000 of solution are given in order that the 

 results may be readily intelligible to readers who are familiar with 

 one or the other method, as the case ma}^ be. 



The method pursued in these experiments was to make and care- 

 fully standardize a large volume of a normal solution of each salt 

 and then dilute to the required strength as occasion demanded. 



In beginning the experiments the limit for each salt as determined 

 by investigators in the field was first tried, but immediately showed 

 itself to be too high. So lower and lower concentrations had to be 

 tested until the critical one was reached. 



PLAINTS SELECTED FOR EXPERIMENT. 



For a variety of reasons the white lupine (Lupiiiiis dlbus) was 

 employed in nearly all the experiments, although subsequently alfalfa 

 (MecUcago sativa) was introduced for comparison. The lupine has a 

 seed of good size, averaging 10 to 12 mm. in greatest diameter. As an 

 abundant supply of nutritive material is stored in the thick seed 

 leaves, there is no danger of starvation of the seedlings in experi- 

 ments of short duration such as those here described. The lupine 

 seeds germinate readily, sending out a vigorous radicle with clean, 



^Dandeno [Bot. Gazette 32, 229 (1901)] has recently called attention to a certain 

 amount of confusion which has existed among- both chemists and physiologists 

 as to the preparation of a normal solution, and it has seemed wise to describe in 

 detail the procedure followed in this investigation. 





f 



