1913] J- Arthur Harris and Ross Aiken G ortner 527 



value of the data been completely destroyed by the reduction of all 

 frequencies to percentages. We must confess that these evidences 

 of gross carelessness in the treatment of their numerical data lessen 

 our confidence in the conclusions they and their reviewers have 

 drawn from it. 



Turn now to three other series of analyses by Andrlik and 

 Urban made for a different purpose.*^ They give the weight, sugar 

 content and nitrogen'^ content of 36 beets grown from normal- 

 nitrogen-content mothers and the same data for 36 roots grown 

 from high-nitrogen-content parents. The first of these we designate 

 as Series A, the second as Series B. They also give 100 analyses 

 of beets from the same parent individual. We refer to these as C. 

 For weight and sugar the correlations are : 



For Series A, r„, = — 0.278 ± 0.104 

 For Series B, r^, = — 0.456 ± 0.089 

 For Series C, r^, = — 0.031 ± 0.067 



All the relationships are negative. The probable errors are high 

 because of the small number of beets given, but the two first cases 

 may perhaps be regarded as statistically significant in comparison 

 with their probable errors. They are in excellent agreement with 

 the correlations published in our first paper. 



We have also deduced the coefficients for weight and nitrogen, n. 

 They are : 



For Series A, r„„ = — 0.234 ±0.106 

 For Series B, r„„ = - 0.393 =*= oo95 

 For Series C, r„„ = + 0.054 =^ 0.067 



One of the constants is positive in sign but less than its probable 

 error. The two negative values are fairly large and possibly sig- 

 nificant in comparison with their probable error. 



Still another series of data^ of quite a different sort affords, it 



8 Andrlik, K. and J. Urban ; "Über die Variabilität des Stickstoffgehaltes in 

 Zuckerrübenwurzeln : Zeitschr. f. Zuckerindustrie in Böhmen, 1912, xxxvi, 



p. 513- 



''■ Milligrams of nitrogen in 100 grams of root. 



8 Novotny, K. ; Ein Beitrag zur Betrachtungen über die Beziehungen 

 zwischen dem percentuellen Zuckergehalte und dem Gewichte der Rüben : 

 Zeitschr. f. Zuckerindustrie in Böhmen, 1912, xxxvi, p. 269. 



