Factors affecting S- methyl cysteine sulphoxide (SMCO) content in Maris Kestrel kale were studied in field and glasshouse trials. SMCO is responsible for kale anaemia in ruminants. Higher plant populations reduced whole plant and stem SMCO in six trials, especially at harvests after October. Increasing plant density from 10 to 80 plants /m2 reduced whole plant SMCO in November or December by approximately 25 %. Nitrogen (N) fertiliser increased SMCO and dry matter (DM) yields in six out of seven trials, with SMCO increases ranging from 8 to 68% depending on soil fertility. There was a high correlation between SMCO (Y) and N (x) levels in mature kale from different sites (r = 0.92, n =26) . Using the relationship Y = 7.53x - 3.20, plant N could be measured as an estimate of the SMCO level. On free draining soils low in available sulphate at sowing (< 12 ppm) applied sulphur (S) increased SMCO in two out of six trials, although only at early autumn harvests. On a soil with impeded drainage (sulphate = 12.5 ppm), S tended to increase (P <0.10) SMCO at a December harvest. As the crop aged, SMCO increased especially in leaves. This was not caused by frosting, but may have been affected by floral initiation, as stems with small inflorescences had higher (P < 0.05) SMCO levels than non -flowering stems. July sown kale had less SMCO than June sown kale in September, but tended to have more SMCO in November, especially at higher populations and particularly in stems. SMCO was not affected by varying growing temperature ( -6 to 15°C) , moisture availability, chopping or wilting for 96 hours. Chrysol (13.1 g /kg DM) tended to have a higher SMCO content than Canson, Maris Kestrel, Merlin, Vulcan or Bittern (mean = 9.2 g /kg DM) . The lower yielding cultivars, Chrysol and Canson, contained most of their SMCO in leaf tissue. SMCO varied in different plant components, with SMCO in petioles having the highest correlation with whole plant SMCO (r = 0.73, n =8 and 18) . Near infrared reflectance (NIR) analysis accurately predicted N and moisture contents in different kale tissue types, with correlation coefficients (r) > 0.95 between NIR predicted and Kjeldahl N or oven-dried moisture values. SMCO was only predicted satisfactorily by NIR for plant breeding purposes in entire leaves and young leaves, which had average standard errors of 2.3 and 1.7 respectively and correlations between SMCO values predicted by NIR and the autoanalysis method of r = 0.87 to 0.91. In stems and whole plants, which had average standard errors of 2.9 and 3.9 respectively and correlations (r) ranging from 0.69 to 0.84, NIR could only be used to separate low and high SMCO samples.