276 



have opportunity for more natural root development. The temperature 

 of the soil ill the boxes is subject to less tluctuatious thau in the small 

 l>ots, and is more like that of the soil in its natural state. The v<m- 

 ditions of <>ro\vth are thus metre nearly normal. With i»roper ai>i»li;iiices 

 the water supply can be kejjt under control, the water taltle maintuined 

 at any desired level, the drainage water collected, measured, a 1 id ;n I :ily zed, 

 and thus the statistics of incctnu* and ont.uo of water and plant food can 

 be deternnned. 



Plans of box experiments adopted in the exjierimental jrarden of the 

 Agricultural Institute of the University of Halle by Dr. Weddtmann, 

 and at the newly orjianized Exju'riment Station for Plant Culture in 

 Dresden by Dr. Stej^lich, are «;iven on jta.ucs .'UL*-3.~)-i of the ]iresent 

 mnnber of the Kecord. 



By this method many of tlie dillicnlties of field e\|>erimentin<i' are 

 a^<>ided, most of its advantajics are retained, and a nundu'r of very 

 desirable features are added. The ex]>eriments are (dose at hand and 

 more easily watched than in a distant field. Any desired soil or selection 

 of soils can be used. Eaidi soil can be earefnlly jireitared, sifted fre«' from 

 stones, and tlKtroUi^ldy nnxed, so tliat it sliall be alik«' in all tlie boxes. 

 Tlie boxes are set in the ground and plants are urown in tlie spaces 

 between them. We thus have a series of «'X]»erimental iilats. asina Held 

 experiment, oidy on a smaller and more convenient scale. Kaidi jdat 

 is entirely isolati'd, so that thei)lants of one jdat can not feed ujion the 

 nniterial of another, and is large enough for a number of plants to grow 

 upon it and small emaigh to l)e mamiged with little labor. There are 

 tubes for watering the soil from below and collecting the drainage water. 

 The water suj)i)ly can be regulated at will and made the same for all. 

 Thus drouth and excessive moistur«», whicdi injure or ruin so many 

 experiments in the field. ar«' i»revente(l. and two chief sonrcesof error in 

 field experiments — unevenness of soil and ine(iuality<d" water sujtply in 

 the (liflereut jdats — are avoided. Data for the statistics of water sui>i)ly 

 .and removal are found in the measurements of rainfall, water a]>itlied 

 artificially, and drainag*'. Analysesof drainage wati'r show the amounts 

 of plant food removed thereby. These, with weights and analyses of 

 soil, fertilizers, and crops, sui^jdy the statistics of gain and loss of the 

 elements of plant fi)od. The i)lants can be more easily cared Ibi- and 

 better protected from depredations of animals and from disease than in 

 field experiments. The quantities of produce are so snmll that the 

 plants and their several parts can be easily measured, and can bo 

 weighed and analyzed with comparatively little labor. In short, this 

 is a sort of field laboratory. It is a device for applying to field «'xperi- 

 menting the accuracy of method and convenience of mani])ulation which 

 are indis])ensable for the best success. 



A method of field experimenting which has a nund)er of the advantages 

 of the box experiment, has been adopted at the Dresden Station and 

 is described on pages 350, 351. Small i>lats are used. These arc made 



