31
SEED TESTING INTERNATIONAL APRIL 2026
• RULES DEVELOPMENT
method was assessed through the estimation 
of repeatability and reproducibility parameters 
in the context of binomial data. A linear mixed 
model (LMM) was fitted using a R function 
developed by the ISTA Statistics Committee; 
statistical analyses were carried out using four 
replicates of 50 and could reveal means out 
of tolerance. The meaningfulness of the cold 
test was also evaluated by correlation analysis 
between cold test, germination test and field 
emergence results.
Results
Establishment of Relationship Between 
Field Emergence and Cold Test
Means of daily temperature were recorded in 
the four locations. Sowing temperature varied 
between 7.8 °C in Chappes (earliest date of 
sowing) and 10.1 °C in Haut-Mauco (sown 25 d  
later). The highest variability of temperature 
was recorded in Chappes and the monthly 
temperature corresponding to the final field 
emergence was 11.1 °C for Arras, 11.5 °C for 
Chappes, 12.3 °C for Saint Mathurin sur Loire and 
13.6 °C for Haut-Mauco.
There was a significant relationship between the 
cold test and field emergence (r ≥ 0.95) for the 
three seed lots sown early (Table 2), especially 
in the northernmost field (Arras with r = 1). 
However, there was also a positive relationship 
(r = 0.96) between standard germination and 
field emergence in the south-west (Haut-Mauco), 
where the six seed lots used for the comparative 
test were considered (Fig. 2).
It is interesting to note the higher correlation 
of field emergence with cold test results and 
the different levels of field emergence for seed 
lots with very similar germination levels. For 
example, lots M1 and M3 with above 90% normal 
germination gave field emergences of 90% and 
60% respectively in the north of France (Arras) 
and had a difference of 9% in the favourable field 
(Haut-Mauco). In this traditional maize crop 
area, the lowest emergence was 82% for lot M5, 
with M1 and M6 having field emergence close to 
their standard germination (Fig. 3).
Nevertheless, the seed lot ranking was very 
similar between laboratory and field (Table 3), 
with M1 and M6 having the highest vigour and 
M5, the lowest vigour.
Comparative Cold Tests Between 
Laboratories
Normal germination counts after cold testing 
revealed differences between seed lots in each 
laboratory (Fig. 4). Results revealed seed lot M5 as 
low vigour and seed lots M1, M2 and M6 as high 
vigour lots between all participating laboratories. 
Overall means comparison ranked the six seed 
lots in three classes (Table 3).
Repeatability and reproducibility estimates 
are presented in Table 4 for all data. Dispersion 
factors were close to 1 for the cold test, indicating 
that there is no evidence of overdispersion within 
laboratories (⨍𝑟 ), nor between laboratories (⨍𝑅).
Figure 2. Relationships between final field emergence in four locations and mean cold test results 
(orange dots) or mean standard germination results (green dots) in controlled conditions; significant 
relationships are indicated with *p>0.05, **p>0.01
M1
M2
M3
M4
M5
M6
Lab cold test (%)
93.6 ± 1.5
91.3 ± 1.8
86.3 ± 3.5
83.6 ± 2.6
72.7 ± 2.3
91.1 ± 3.6
Haut-Mauco (%)
98.0 ± 1.9
89.0 ± 3.4
89.0 ± 3.4
88.0 ± 2.2
82.0 ± 5.6
95.0 ± 3.7
Arras (%)
90.0 ± 5.5
82.0 ± 4.6
65.0 ± 5.8
nd
nd
nd
Chappes (%)
93.0 ± 6.7
83.0 ± 10
70.0 ± 6.8
nd
nd
nd
St Mathurin (%)
97.0 ± 1.4
93.0 ± 2.5
88.0 ± 2.4
nd
nd
nd
Table 2. Mean results (±SD) for final field emergence (28 d after sowing) in four locations and mean 
cold test results (±SD) obtained in six laboratories
Figure 3. Mean normal germination after 12 d cold test and final mean field emergence (%) of the 
samples of Zea mays in the four experimental plots (4×100 seeds); lots M4 to M6 arrived too late to 
be sown in the first three places
M1
M2
M3
M4
M5
M6
7 d @ 10 °C – Sand
93.6a
91.3a
86.3b
83.6b
72.7c
91.1a
Table 3. Cold germination means (%) for six seed lots of maize assessed after 7 d at 10 °C and 5 d at 
25 °C in six laboratories; each mean is derived from four replicates of 50 seeds per laboratory; for 
each method, means with different letters are significantly different according to Tukey test (α=0.05)

View this content as a flipbook by clicking here.