The transition towards more sustainable plant protection with reduced pesticide use is difficult, because there is no ‘silver bullet’ available among non-chemical tools. Integrating several plant protection approaches may thus be needed for efficient pest management. Recently, increasing the genetic diversity of plantations via cultivar mixing has been proposed as a possible method to reduce pest damage. However, previous studies have not addressed either the relative efficiency of exploiting cultivar mixing and intrinsic plant herbivore resistance or the potential utility of combining these approaches to increase cropping security. Here, using a full factorial experiment with 60 woodland strawberry plots, we tested for the relative and combined effect of cultivar mixing and intrinsic plant resistance on herbivore damage and yield. The experiment comprised two levels of diversity (“high” with ten varieties and “low” with two varieties), and three levels of resistance (“resistant” comprising only varieties intrinsically resistant against strawberry leaf beetle Galerucella tenella; “susceptible” with susceptible varieties only; and “resistance mixtures” with 50:50 mixtures of resistant and susceptible varieties). The experiment was carried out over two growing seasons. Use of resistant varieties either alone or intermixed with susceptible varieties in “resistance mixtures” reduced insect herbivory. Interestingly, resistant varieties not only reduced the mean damage in “resistance mixtures” by themselves being less damaged, but also protected intermixed susceptible varieties via associational resistance. The effect of higher genetic diversity was less evident, reducing herbivory only at the highest level of herbivore damage. In general, herbivory was lowest in plots with high diversity that included at least some resistant varieties, and highest in low diversity plots consisting only of susceptible varieties. Despite this, no significant difference in yield (fruit biomass) was found, indicating that strawberry may be relatively tolerant. Our results demonstrate that combined use of high genetic diversity and resistant varieties can help reduce pest damage and provides a useful tool for sustainable food production. “Resistance mixtures” may be particularly useful for sensitive food crops where susceptible varieties are high-yielding that could not be completely replaced by resistant ones.

Field experiment covering a total area of 130 × 40 m (i.e., similar to the general size of strawberry plantation) was established at the Alnarp campus of the Swedish University of Agricultural Sciences in southern Sweden.Leaf damage data  was collected four times during 2018 and once in 2019 from woodland strawberry (Fragaria vesca) plants in a full factory experimental design consisting of two levels of genetic diveristy (i.e. number of varieties used, 10= high and 2 = low) and three levels of intrinsic plant resistance on the experimental plots (R= resistant against a generalists leaf chewing herbivore, Galerucella tenella, S= susceptible against this pest and M= resistance mixture plot, consisting of 50:50 ratio of both resistant and susceptible varieties).The full factorial set up resulted in the following treatment combinations: high diversity resistant (HR), high diversity susceptible (HS), high diversity mixture (HM), low diversity resistant (LR), low diversity susceptible (LS) and low diversity mixed (LM) plots, each of which was replicated ten times. The position of the varieties within plots were randomized. Each plot contained forty plants, and two replicates of each variety were grown in the same soil bag, resulting in 2400 plants in total.

Leaf damage was observed as number of damaged leaf (regardless of the size of the damaged area). Plant size (photosynthetic area) was recorded at the last scoring occasion in 2018, measured as height x width of the plant. Position of the plot in a field was categorized into 4 categories to take into account potential edge effects: (top = 'forest' = the 4 plots in the first row;  bottom = 'end' = the last row i.e. 4 plots ;  right = 'edge right' =1 vertical colum of plots; left = 'hedge' = 1 vertical colum of plots; and 'middle'= two vertical colums in the middle of the field). 

In order to investigate the effect of plot’s resistance and diversity on yield, we harvested all ripe fruits on a weekly basis during May and June during 2019. The fruits were weighed on the day of harvest. For practical reasons, the fruits from four plants of the same variety were pooled and weighed resulting in ten samples per plot. Yield data were also collected during 2018. However, due to the natural spatiotemporal variation in herbivory, the arrival of herbivores in 2018 occurred after or during the fruit harvest. Therefore, the yield data from 2018 were not analyzed in relation to herbivory.

For statistical models, we calculated the average yield per plant by dividing the weight of the sample box by four (as samples were harvested by combining fruits from four plants of the same variety) to make the dataset comparable to leaf damage analyses. Similarly, when using plant size as a covariate in yield analyses, we first calculated the average plant size for each variety in a plot and divided it by four to make the plant size data comparable to yield data.

The plant damage was scored 4 times in 2018, but due to the very low level of herbivory in the first scoring occasions (obs time1), this time point was not analysed in the publication.  Yield data is only from year 2019, as  due to the natural spatiotemporal variation in herbivory, the arrival of herbivores in 2018 occurred after or during the fruit harvest  For practical reasons, the fruits from four plants of the same variety were pooled and weighed resulting in ten samples per plot. 

For statistical models, we calculated the average yield per plant by dividing the weight of the sample box by four (as samples were harvested by combining fruits from four plants of the same variety) to make the dataset comparable to leaf damage analyses. Similarly, when using plant size as a covariate in yield analyses, we first calculated the average plant size for each variety in a plot and divided it by four to make the plant size data comparable to yield data.

For leaf damage data

Year= year of sampling
running number= running number of the row
edge= position of the plot in the field
block= id number of the block
plot= id number of the plot
bag= id number of the soil bag where 2 plants of the same genotypes were grown
resistance= resistance of the plot, i.e. if the plot contained only resistant or susceptible plants or their 50:50 mixture
diversity= high = 10 genotypes, low = 2 genotypes
plot2= id code for the plot including the level of diversity and resistance
plant_res= plant resistance, resist or susceptible against the strawberry leaf beetle, Galerucella tenella
observation_time= date of the data collection
obs_time= number for scoring occasion
period= time in the growing season when leaf damage was estimated
genotype= genotype id of the plant
plant id= code for plant id, including plot and genotype and id number
plant id2= more comprehensive id code for a plant, including block, genotype and plant number
damage leaves= number of insect damaged (leaf chewing marks) on the plant
plant area= plant size calculated at the end of the growing season in 2018, obs_time 4. Area describes the photosynthethic area height x width cm of the plant (cm2)

For yield data

edge= position of the plot in the field (same as above)
resistance= resistance of the plot, i.e. if the plot contained only resistant or susceptible plants or their 50:50 mixture
diversity= high = 10 genotypes, low = 2 genotypes
block= id number of the block
plot= id code for the plot including the level of diversity and resistance
genotype= id code for genotype (i.e. variety)
weight 1-4= for pracitcal reasons, woodland strawberry fruits from the same genotype were collected in maximum 5 plastic boxes, these columns show the weight of each box (g)
total weight= summarized weight per genotype from all the fruits
area2018= plant photosynthetic area (size) measured at the end of previous growing season in 2018, measured as plant height x widht
average_weight_plant= average yield per plant calculated by dividing the total yield weight by 4 (i.e. the number of plants per genotype from which the fruits were collected together)
average_area= average photosynthetic area for plant, calculated by the total area per genotype, divided by 4 (number of plants where fruits were collected together)