Ecological restoration frequently involves the addition of native plants, but the effectiveness (in terms of plant growth, plant survival, and cost) of using seeds versus container plants has not been studied in many plant communities. It is also not known if plant success would vary by species or based on functional traits. To answer these questions, we added several shrub species to a coastal sage scrub restoration site as seeds or as seedlings in a randomized block design. We measured percent cover, density, species richness, size, survival, and costs. Over the two years of the study, shrubs added to the site as seeds grew more and continued to have greater density than plants added from containers. Seeded plots also had greater native species richness than planted plots. However, shrubs from containers had higher survival rates, and percent cover was comparable between the planted and seeded treatments. Responses varied by species depending on functional traits, with deep-rooted evergreen species establishing better from container plants.  Our cost analysis showed that it is more expensive to use container plants than seed, with most of the costs attributed to labor and supplies needed to grow plants. Our measurements of shrub density, survival, species richness, and growth in two years in our experimental plots lead us to conclude that coastal sage scrub restoration with seeds is optimal for increasing density and species richness with limited funds, yet the addition of some species from container plants may be necessary if key species are desired as part of the project objectives.

Plant density data were collected in April and May of 2017 and 2018. We randomly placed a 50 cm by 50 cm quadrat near the center of each plot and then counted all native and non-native species rooted within the quadrat. The native/non-native density and species richness were calculated based on the species identity and number of plants found in each quadrat.

To measure percent cover, we installed five 4 m vegetation sampling transects, spaced 0.5 m apart from each other with point data collected every 1 m to reach a total of 25 points per plot. At each point, we recorded all species present. The raw transect data has been rearranged to list each plant (or ground cover) in its own row.

To evaluate the differences in overall plant health between seeded and planted individuals, we flagged individuals in both treatments and recorded survivorship over time. We collected plant size measurements on flagged native shrubs, measuring height, width, and length. Shrubs were checked on several different dates and either measured or noted as dead. We used these data to calculate the average lifespan (the number of days between the estimated date of germination and the date of measurement).

We used size measurements from 5/26/18 to compare final size, and the change in size from 5/30/17 to 5/26/18 to compare growth. Relative growth was calculated as the relative change in size (height or width) using the formula: (size on 5/26/18 − size on 5/30/17) ∕ size on 5/30/17.

We kept track of the hours of labor spent during the first year of our study (2016-2017) to compare differences in the time devoted to seeding or planting. We also kept a record of all relevant expenses related to the project.

The data from block 5 has been removed from all datasets. This block was overrun with non-native invasive plants and yielded no useful information.
Dead individuals were always excluded from all datasets.
Blocks A-G (the “New” blocks) were not used in our analyses due to the very poor survivorship of both seeded and planted plants in all plots.