Amcrlcini Fisltrrirs Socicli/. 91 



the to}) of tlu'iii so that tlioy will iloat about level with the sur- 

 face of the lake or slightly submerovd. In some of these boxes 

 Ave will make a deposit of soil six inches or more dee]) This soil 

 is to be taken from the lake bottom in some localitv wliei-e there 

 is an abundant growth of plants. In this we will plant a cer- 

 tain number of individuals of uniform size, of some one species 

 that have heen carefully selected from specimens growing in the 

 lake. In the other hox we will attach a like munbt'r of such 

 ])lants to wooden bars, and fasten these bars so tliat the plants in 

 the two boxes are in the same depth of water. AVe thus have for 

 comparison, two grou})s of ])lants, one of Avhich is surrounded 

 by natural conditions, and the other has natural conditions ex- 

 cept that the soil is absent, and the plants can get only such 

 ]iourishment as is provided !)y the water. After a i)eriod of four 

 weeks very marked differences between the two groups of plants 

 may be noted. The plants rooted in soil look as strong and 

 healthy as tliose growing in the lake. The volume of vegetation 

 produced Ijy the former group was twice that produced by the 

 latter group, while the suspended plants were stunted in growth 

 and manifested the ordinary signs of unfavorable environment. 

 If we now collect our two groups of ])lants, and after carefully 

 washing, obtain the dry weight, it will be found to be a third 

 more for the plants rooted in soil. It is thus evident, that soil is 

 necessary for the best growth of these plants. If now we com- 

 pare equal volumes of the fresh plants in our experiment, it will 

 appear that the dry weight of the suspended plants is greater 

 than that of those rooted in soil. This must mean that dill'erent 

 ])hysiological processes have been operating in tlie two groups. 

 If a microscopic examination of the fresh plants be made, it will 

 be found that the tissues of the plants rooted in soil contain rela- 

 tively little starch, while those of the suspended ])lants are liter- 

 ally gorged with starch. Thus is explained our discovery con- 

 cerning the dry weight of equal volumes. It is the abnormal in- 

 crease in the amount of starch, that makes the sus])ended ])lants 

 weigh more. If now we compare the two groujis with regard to 

 their chemical composition, we shall find that the plants which 

 were denied the soil, contain a smaller proportion of nitrogen, of 

 ])otash, and of phosphoric acid. To recall now the results of 

 this line of investigation, we mav sav that when the plants are 



