Many habitat-specialist plant species have been endangered by the fragmentation or destruction of their habitats. We evaluated two conservation methods to reinforce M. smejkalii populations: sowing seeds collected from natural populations and transplanting pre-grown individuals in three consecutive years to sites currently or recently occupied by the species. We checked for seed emergence (recruitment) and followed the survival and performance of the transplants and recruits to identify factors responsible for the establishment success for up to three years after transplantation. Recruitment from sown seeds was unsuccessful, but overall, 83% of transplants survived the first year and 49% until the third year. Survival depended on the year of transplantation, the size of the transplanted individual, and the mother plant's characteristics. Mother plant characteristics were important for growth and flowering in the first year, but not for survival. Across years, initial size and age were more important for the size and number of flowers than mother plant identity. Population differences had a larger effect on the size and number of flowers in the first year than across years. Seedlings established from seeds of the transplants in four of the five populations. This led, despite the mortality of transplants, to a four to six times larger population size after three years in two populations, compared to the number of transplanted individuals.

Measurements

The length, width, and the number of flowering and non-flowering stems of each mother plant (which served as seed source) were measured in June in natural sites, and mother plant size was calculated as the area of an ellipse (because M. smejkalii grows predominantly vertically in separate stems, area is a better fitness estimate than plant height). We also counted the number of flowers on ten randomly chosen stems with terminated growth (dichotomous stem division) and multiplied this by the number of flowering stems to calculate the total number of flowers for each flowering plant. The number of flowers per individual stem was averaged per population and year. Prior to transplantation, we measured the maximum length and width of each plant to calculate plant size as above and determined if the plant was flowering or not. Plant characteristics in the years after transplantation were evaluated likewise at the time of flowering during June and July. The data collection was done at two levels: whole site and transects (permanent plots).

At the level of the whole site, we evaluated the status of each transplanted individual in the following years as flowering, non-flowering, dry, absent (but potentially still able to regenerate in the following years), and lost (when both plant and label could not be found). Survival was encoded as 0 or 1. Individuals were counted as alive (=1) if they could be found, even though they were dry at the time of observation. Individuals that were absent in one or two years but re-emerged in the next year were also counted as alive in the years they had been missing previously (0.26 % of total individuals). Absent individuals (no aboveground parts visible or disappeared for unknown reasons, including the tag) and did not re-grow later were considered dead (=0). Additionally, we counted the total number of newly recruited plants at the whole part of the site, where the transplantation had been done.

Transects with a width of two meters passed through the area of newly transplanted individuals and covered the heterogeneity of the environment, and at least 50% of the transplanted individuals, to reduce sampling effort in the years following transplantations. The transect length and number of transects per population were dependent on population size and plant distribution in the terrain. In these transects, we measured and calculated the same parameters as above (size,  flowering, and non-flowering stems and flowers).

Recruits within the transects in the years after transplantation were numbered and tagged, and similarly recorded and measured in consecutive years as the transplanted individuals.

The relative growth rate (RGR) of individual plants was computed for each measurement time (tx, with x being 1, 2, or 3 years since transplantation) compared to the initial measurement before transplantation (t0)

RGR = [ln(Size_tx) - ln(Size_t0)] / [tx-t0]