Germination behaviour of Conyza bonariensis to constant and alternate temperatures across different populations

Conyza bonariensis is one of the most problematic weed species throughout the world. It is considered highly noxious due to its interference with human activities, and especially the competition it poses with economically important crops. This research investigated the temperature requirements for seed germination of four populations of C. bonariensis with distinct origin and the influence of daily alternate temperatures. For this, a set of germination tests were performed in growth chambers to explore the effect of constant and alternate temperatures. Seeds of the four populations (from Lleida, Badajoz and Seville, Spain and Bahía Blanca, Argentina) were maintained at constant temperatures ranging from 5–35oC. The final germination and cardinal temperatures (base, optimum and maximum) of each population were obtained. We also tested the influence of daily alternate temperatures on final germination. To do so, seeds were exposed to two temperature regimes: 5/15, 10/20, 15/25, 20/30 and 25/35oC night/day temperature (intervals increasing 5oC, with constant oscillation of 10oC) and to 18/22, 26/24, 14/26, 12/28 and 10/30oC night/day temperature (intervals with average of 20oC, but increasing the oscillation in 4oC between intervals). In general, all populations behaved similarly, with the highest germination percentages occurring in the optimum temperature range (between 21.7oC and 22.3oC) for both constant and alternate temperatures. In general, climatic origin affected germination response, where seeds obtained from the coldest origin exhibited the highest germination percentage at the lowest temperature assayed. In addition, we observed that the alternate temperatures can positively affect total germination, especially in oscillations that were further from the average optimum temperature (20oC), with high germination percentage for the oscillations of 15/25, 20/30, 18/22, 16/24, 14/26, 12/28 and 10/30oC in all populations. A cc ep te d A rti cl e This article is protected by copyright. All rights reserved. The cardinal temperatures obtained were significantly different across the populations. These results provide information that will facilitate a better understanding of the behaviour of Conyza and improve current field emergence models.

etc.) in several countries (Argentina, Australia, United States or Spain), in addition to evolving multiple resistance (PSI Electron Diverter and EPSP synthase inhibitors) (Heap, 2019).
Conyza is photoblastic, emerging from the upper layers of the soil surface (0-2 cm) with limited persistence, as it has very low dormancy levels and the viability of the ungerminated seeds is severely lost in the first year (Wu et al., 2007). Annual weed species survival is highly dependent on seedling emergence and recruitment (Forcella et al., 2000). Thus, it is important to know both timing and magnitude of seedling emergence in the field in order to implement successful control measures for weeds (García et al., 2013;Royo-Esnal et al., 2015). In this respect, Zambrano-Navea et al.
(2013) modelled the emergence of C. bonariensis and developed a cohort-based stochastic model of the population dynamics (Zambrano-Navea et al., 2016). However, studying the germination response of more populations and at more temperatures and intervals would expand upon these existing models. Conyza bonariensis biology is well understood, but additional information regarding germination temperature thresholds is required to implement integrated management control measures. An added complexity is that the variation of threshold parameters between populations can be significant due to local adaptations (Tozzi et al., 2014;Bajwa et al., 2016). For example, in C. bonariensis, Karlsson & Milberg (2007)

Accepted Article
This article is protected by copyright. All rights reserved.
temperatures and at different alternate temperatures. An additional objective was to compare the cardinal temperatures (T b , T o , T c ) of each population.

| Plant material collection
Conyza bonariensis seeds were harvested at maturity in September 2016 in Spain and in November 2017 in Argentina. Seeds from Spain were collected from three different habitats: a vineyard in Lleida (41.658010, 0.523766), a garden in Seville (37.352824, -5.933194) and an olive orchard in Badajoz (38.702537, -5.573246). The population from Argentina belonged to a garden in Bahía Blanca (-38.695394, -62.253302). The four locations have specific climatic conditions (Table 1). According with Torra et al.
(2016), seeds were collected from different plants throughout the field, were air-dried under laboratory conditions for one week and dry stored in the dark in paper bags at 4ºC until the beginning of the experiment.

| Experimental design
Three germination tests were established at different temperature conditions and repeated twice: the first one at constant temperatures, the second at constant day/night temperature oscillations (T osc ) and different mean temperatures (T m ), and the third one at different day/night T osc but with the same T m . All the experiments were performed at the Departamento de Agronomía, Universidad Nacional del Sur and CONICET (Bahía Blanca, Argentina). In all three tests, batches of 30 seeds were sown on 9-cm Petri dishes lined with a Nº1 filter paper layer wetted with distilled water. Four replicates per population and temperature were used following a completely randomized design.

Accepted Article
This article is protected by copyright. All rights reserved.
Germinated seeds were counted on a daily basis until 21 days or until no further germination occurred during five consecutive days.
For the different tests, a seed was considered germinated when the radicle had extended more than 1mm beyond the seed coat (Steinmaus et al., 2000;Wu et al., 2007).

Accepted Article
This article is protected by copyright. All rights reserved.
Germinated seeds were removed from the dish once counted. Seed viability at the end of the germination tests was assessed by counting the number of germinated seeds after incubation at 20ºC (Wu et al., 2007) under a 12-h photoperiod for five days.

| Statistical analysis
Total germination percentages between populations and incubation temperatures as well as cardinal temperatures were subjected to analysis of variance (ANOVA). The SED and LSD are provided.

Eq. [1]
Where Y is the germination percentage, a is the maximum germination percentage, d 50 is the time in days to achieve 50% of germination and b is the germination rate at d 50 .
Estimation of the optimum temperature (T o ): Once d 50 was defined, its inverse value (1/d 50 ) where represented in a figure and a threeparameter Lorentzian function was fitted, equally, to each replicate of each population (Eq. [2]).

Eq. [2]
Where Y is 1/d 50 value at each temperature x, a is the maximum1/d 50 value, x 0 is the temperature at which the highest value of 1/d 50 is obtained, and coincides with the

Accepted Article
This article is protected by copyright. All rights reserved.
centre of the peak and the optimum temperature at the same time; and parameter b is the mean width of the peak.
T b and T c estimation: Once T o was defined, sub-optimal temperatures were used to obtain T b and supraoptimal temperatures were used to obtain T c , and regression lines were fitted, respectively, to each (Eq.

Eq. [3]
Where Y is the 1/d 50 value at each temperature x, a is the slope and b is a constant value.

Accepted Article
This article is protected by copyright. All rights reserved.
At constant temperatures, the highest germination percentages for all populations were obtained between 15 ºC and 25ºC ( With respect to test 2, the seed exposure to a constant T osc and at different T m significantly affected the germination percentage. Statistical differences were found between populations and incubation temperatures at all constant temperatures (P <

Accepted Article
This article is protected by copyright. All rights reserved. 0.001) except for 20-30ºC (P < 0.397). Higher germination percentages were observed at 20-30ºC for populations from Lleida and Seville, 5-15ºC for Badajoz and 15-25ºC for Bahía Blanca, with germination percentages higher than 94% for all of them (Test 2, In the case of test 3, where a same T m and different T osc was assessed, no significant differences in germination percentage were observed inside each population (P = 0.327 for Lleida, P = 0.780 for Badajoz, P = 0.334 for Seville and P = 0.09 for Bahía Blanca), but there were differences between the different T osc considered (P < 0.001), (Test 3, Table 3). At all temperatures assayed, significantly lower germination percentages were observed between populations from Seville (with percentages between 71.7% and 85.07%) and the rest of the populations.
The effect of temperature did not only affect the final germination percentage, but also the germination timing and rate ( Figure 1). The cumulate germination of all populations at all temperatures successfully fitted to log-logistic function, except for those temperatures at which germination was too low or null (Table 4). At the lowest and highest constant temperatures, lower values were estimated for parameter x o , indicating a delay in germination (Table 4; Figure 1). This behaviour is similar in all the populations. In general, the germination rate, identified as parameter b (Table 4) was

Accepted Article
This article is protected by copyright. All rights reserved.
faster between 15ºC and 25ºC, compared to at 5ºC, 10ºC and 30ºC, except for the population from Badajoz. Parameter b could not be significantly fitted (P<0.05) for populations from Lleida and Bahía Blanca at 20 ºC and 25ºC, due to the fast germination rate, though the log-logistic function was significantly fitted (Table 4).  (Table 4).

| Estimation of T b , T o and T
The lowest T c value was also obtained from the Bahía Blanca population (31.5ºC), followed by that from Seville (31.7ºC) and Lleida (32.3ºC). Finally, the highest value was observed in the Badajoz population (34.0ºC). Statistical differences were found between populations (P < 0.001) for T b but not for T o and T c (P < 0.246 and P < 0.103, respectively).

Accepted Article
This article is protected by copyright. All rights reserved.
All populations showed similar germination behaviour. Germination percentage was highest near the optimum temperature and there were significant differences in the final germination percentages, which appeared mainly at lower temperatures. These differences could be explained, in part, by the climate of the original localities, but also by a possible maternal effect. The environmental conditions under which the mother plant produced the seeds, and also the position of the seed in the plant can impact seed germination. Likewise, water deficit, the age of the plant, the day length, the parental photo-thermal environment, light quality, altitude, and temperature are known, among other factors, to affect germinability (and dormancy in some cases) in other species (Gutterman, 2000;Menegat et al., 2018).

| Effect of temperature on germination
For all populations, the maximum germination percentage was reached near 20ºC. The estimation of the optimal temperature (T o ) (21.7ºC to 22.3ºC) allowed for little distinction between populations (Table 4). When the temperatures moved away from the optimal, the final germination percentage decreased. This decrease was faster for supraoptimal than for sub-optimal temperatures (

Accepted Article
This article is protected by copyright. All rights reserved.
Differences in germination percentages are accentuated between populations ( Table 2).
The population from Lleida, which is a comparatively colder location (Table 1)

Accepted Article
This article is protected by copyright. All rights reserved.
Our results also agree with those from Karlsson & Milberg (2007)

| Germination patterns and threshold values
The germination patterns of all population at all constant temperatures were, in general, successfully fitted to a log-logistic sigmoidal function (Table 3, Figure 1). The lack of this adjustment in some cases (Table 3)

Accepted Article
This article is protected by copyright. All rights reserved.
the lowest value (4.9ºC) followed by Bahía Blanca (6.9ºC), which agrees with their local climatic origin. In accordance with this, the T b in Seville (8.9ºC) and Badajoz  Karlsson & Milberg, 2007). In addition to genetic origin, the maternal effect is another factor which could have enhanced differences between local populations.
In our study, there were four degrees (ºC) of difference in T b between the seeds from different origins, thereby impeding the development of a common model. In order to develop a model that could be widely applied, the next step is to test differences in the base parameters and germination behaviour of populations coming from different geographical sites, but belonging to the same climatic biotype. If there are not any differences between them, a more precise model could be developed or the current one created by Zambrano-Navea et al. (2013) could be readjusted to the populations of a certain climatic area.

| Conclusion
The germination percentage of C. bonariensis was higher when close to the optimal temperature obtained (22ºC), both for constant and alternate temperatures. In the intervals with same T osc and different T m , some obtained higher germination

Accepted Article
This article is protected by copyright. All rights reserved.

Accepted Article
This article is protected by copyright. All rights reserved.

Accepted Article
This article is protected by copyright. All rights reserved. SED, standard error of the difference between two means; LSD, least significant difference between two means at P = 0.05; d.f., degrees of freedom associated with LSDs and SEDs.