Previous PageTable Of ContentsNext Page

Can Canopy Temperature Depression measurements help breeders in selecting for yield in wheat under irrigated production conditions?

Maarten Van Ginkel, Matthew Reynolds, Richard Trethowan and Eduardo Hernandez

International Maize and Wheat Improvement Center (CIMMYT), Mexico.


The effectiveness under well-watered conditions of augmenting visual selection by breeders with early generation measurements of Canopy Temperature Depression (CTD) on final yield in wheat was studied in five crosses. While results varied by cross, it was shown that indeed such early selection for cool canopies did complement visual selection. The two-pronged approach of the breeder complementing visual selection with searching for cooler canopies identified almost three times as many high yielding lines. In two crosses, the top lines were derived from the combined method, while in the remaining three crosses almost equal numbers of high-yielding lines were identified with both methods. Overall, CTD was relatively highly correlated with yield under these more optimum conditions (0.74). Clearly CTD has the potential to help breeders in early generations identify those populations that later will give the highest yielding lines.

Media summary heading

Wheat plants with cooler leaves give higher yields even under irrigated conditions. Breeders can use this simple technique to increase efficiency during selection.

Key words

Breeding, physiology, genetic gain.


Canopy temperature depression (CTD) has been shown to be correlated with yield in wheat under drought stress (Blum, 1988) and hot, irrigated conditions (Reynolds et al., 1994; Amani et al., 1996). The value of CTD under non-stressed conditions is less well established. The objective of this study was to assess whether CTD could be used to complement breeder-selection under well-watered, non-stressed conditions.

Materials and Methods

Five crosses were made among four parents in the CIMMYT bread wheat breeding program: Attila, Babax, Borlaug F95 and Lucero-M. The resulting segregating generations were selected by shuttling materials between two contrasting sites in Mexico; irrigated Cd. Obregon in north-eastern Mexico and Toluca in the central Mexican highlands, experiencing high rainfall. The modified pedigree method was followed (Rajaram et al., 2002; van Ginkel et al., 2002). The ‘breeder’ (first author) visually selected the best individual plants in the F2 and these were promoted to the F3 as individual plots. Again the ‘breeder’ visually selected among the F3 plots and bulked seed of these plots was planted in F4 plots. In the F4, CTD measurements were taken on individual plots. The breeder also selected visually the plots with desirable agronomic plant types. Thus two separate germplasm streams were created: ‘Breeder-only’ and ‘Breeder+CTD’. Populations were then promoted to the F5 and individual heads selected based on the breeder’s visual evaluation for purified seed multiplication in the F6. In addition, non-selected bulks of the crosses were maintained. In the F7 (1999-2000 crop cycle) and F8 (2000-2001 crop cycle) replicated yield trials were carried out at Cd. Obregon under fully irrigated and optimally fertilized conditions. The trials compared three selection methods: ‘Breeder-only’, ‘Breeder+CTD’, and ‘bulk’. The number of entries that were finally tested in yield trials was not the same for each method (See Table 1), since the aim was to test the usefulness of CTD in an ongoing breeding program. Hence the breeder simply selected the best lines in each cross, in no way biasing the selection to arrive at equivalent numbers in the end. From the bulk populations in the F5, 10 randomly harvested spikes were harvested and multiplied in the F6, followed by randomly five each being selected for each cross, giving a total of 25. Plots 3m in length consisted of four rows in total, planted two each on two beds of 80cm in width. A latinized alpha-lattice design with two replications was employed. The entire plot was harvested and yield calculated at 12% moisture. CTD was measured on all yield trial plots. Analyses were carried out using GLM (SAS, 2002).

Results and Discussion

The overall genetic correlation of CTD with yield taken during the yield trial phase was 0.74 (P = 0.001). This is relatively high, reflecting 50% of variation in yield being explained by cooler canopies.

The breeder-only selection proved significantly superior to the ‘bulk’ across all crosses (Table 1). In each year, the ranking of the selection methods from low to high yield was ‘bulk’, ‘breeder+CTD’ and ‘breeder-only’. However, the ‘CTD’ method was never significantly superior to ‘bulk’ across all crosses. Across crosses the ‘breeder+CTD’ and ‘breeder-only’ methods were not significantly different, but the ‘breeder+CTD’ did identify almost three times as many equally well yielding lines (154) as ‘breeder-only’ (57). That in and of itself is an interesting finding, since it also indicates that more genetic diversity may have been maintained by including CTD measurements in the selection process.

Table 1. Comparison of three selection methods as expressed in yield (kg/ha) over two years (1999-2001) across five elite crosses.


Mean Yield (kg/ha)


Tukey grouping














When crosses were compared individually, ‘Breeder+CTD’ was significantly superior to ‘bulk’ and to ‘Breeder-only’ in one cross, thus indicating that use of CTD in the F4 in this cross later resulted in lines that out-yielded non-selected and breeder-only selected entries.

However, breeding is not about mean performance across crosses: it aims to identify significant outliers and hopefully many of them within a cross. So more relevant than average yield either across all crosses or by cross, is whether individual lines within a cross could be identified in either the ‘Breeder-only’ or ‘Breeder+CTD’ methodology that yielded more than the best lines derived through the other methods. Indeed, in two out of five crosses, individual lines developed using ‘Breeder+CTD’ out-yielded the top ‘Breeder-only’ entry(ies). In one cross, the top 24 lines were all derived from the ‘Breeder+CTD’ method. Figure 1 depicts the second cross, where four lines from the ‘Breeder+CTD’ out-yielded the best line from ‘Breeder-only’. The highest yielding ‘Breeder+CTD’ line yielded 5% more than the top ‘Breeder-only’ entry. In the remaining three crosses, almost equal numbers of lines among the top ten yielding lines were derived from either the ‘Breeder+CTD’ or ‘Breeder-only’ method.

Figure 1a, b, c. Frequency distribution for yield of lines developed with three selection methods: ‘bulk’, ‘Breeder-only’, and ‘Breeder+CTD’.


The following conclusions can be drawn from this study:

1. CTD is relatively highly correlated to yield under optimum production conditions (0.74).

2. While, on average, across all crosses selection by breeder-only appeared superior to breeder selection augmented by early generation CTD measurements, studying individual crosses highlighted the clearly positive impact of the breeder complementing visual selection with CTD measurements.

3. Selection by breeder complemented by CTD measurements overall identified almost three times as many equally well yielding lines as the breeder alone. This may reflect a larger amount of genetic diversity having been sampled, a positive trend.

4. In one cross, the average yield of all lines identified with breeder selection augmented with CTD measurements was higher than that in the other methods. In two crosses the top-yielding 24 and 3 lines respectively derived from using the ‘Breeder+CTD’ methodology. In the other three crosses a more or less equal number of lines in the top 10 derived from each of the two methods, ‘Breeder+CTD’ or ‘Breeder-only’

5. Clearly under well-watered, optimally fertilized conditions early generation CTD measurements were very useful in augmenting visual selection by the breeder, but results will vary by cross.

6. We recommend that this technology of identifying plants with cooler canopies with the aim to raising yield also under non-stressed conditions is further studied by others.


Amani, I., Fischer, R.A., Reynolds, M.P., 1996. Canopy temperature depression association with yield of irrigated spring wheat cultivars in a hot climate. J. Agron. Crop Sci. 176:119-129.

Blum, A. 1988. Plant Breeding for Stress Environments. CRC Press, Boca Raton, FL, p 72.

Rajaram, S., N.E. Borlaug and M. van Ginkel. 2002. CIMMYT international wheat breeding. Pp. 103-117 in: Bread Wheat Improvement and Production. B.C. Curtis, S. Rajaram, and H. Gomez Macpherson (eds.). Plant Production and Protection Series No. 30. FAO, Rome 2002.

Reynolds, M.P., Balota, M., Delgado, M.I.B., Amani, I., Fischer, R.A., 1994. Physiological and morphological traits associated with spring wheat yield under hot, irrigated conditions. Aust. J. Plant Physiol. 21:717-730.

SAS 2002, Version 8.2. SAS institute Inc., Cary, NC, USA.

Van Ginkel, M., R. Trethowan, K. Ammar, Jiankang Wang, and M. Lillemo. 2002. Guide to Bread Wheat Breeding at CIMMYT. CIMMYT Wheat Special Report No. 5. (revised edition) 52 pp. CIMMYT. Mexico, D.F.: Mexico.

Previous PageTop Of PageNext Page