Table 1. Relationships between ratio of P. nodorum to M. graminicola DNA (log_e-transformed) in harvest samples from the Broadbalk experiment, 1844–2003, and agronomic or environmental factors suggested as explanations for changes in prevalence of the two septoria diseases in England and Wales in the early 1980s.

Factor (ref.)	df	P	R2
August sunshine before sowing (20)*	1	NS	0
Sowing date in autumn (17)†	1	0.14	0.4
Use of fungicide seed treatment†	1	NS	0
December-February temperature, seedling stage*	1	0.03	6.2
Use of foliar fungicide, spring (8)†	1	0.06	2
May-June rainfall, adult plant stage (18)*	1	NS	0
Crop height at harvest†	1	0.07	1.7
Harvest technique†	2	0.001	25
Wheat cultivar (7)†	13	0.06	44
England and Wales area sown to wheat (21)‡	1	0.001	11
Interpolated SO2 deposition (24, 25)	1	0.001	61

Relationships were assessed by linear regression (percentage of variance accounted for R²). NS, not significant.

^*All meteorological data are from the Electronic Rothamsted Meteorological Database (1879-2003) or Central England Temperature Database at the Hadley Centre for Climate Prediction and Research, Bracknell, United Kingdom (autumn 1843-1878; adjusted for systematic differences from Rothamsted).

^†Data are from Broadbalk “white book” record series. See Fig. 2 for more details.

^‡Data are from the Department for Environment, Food and Rural Affairs, Economics and Statistics Department.