In a smaller garden we’d be wise to choose plants that are generally behaved clumpers -- they won’t spread too much by runners or sowing. But even site conditions can affect these plants, as clay soil and dense, layered vegetation will inhibit plant reproduction (in general) while loamy or sandy soil with less plant competition will encourage it (in general). If the majority of plants that are best suited to your site tend to have aggressive natures, it’s best to use ALL aggressive plant species so they butt heads, collide, and help keep each other in check.
We also shouldn’t place plants based on their mature size. If a plant tag, aka thorough internet and book research, shows a full grown plants gets 3’ wide we should still plant it 12” from its neighbor. Why would you do that, especially when plants cost money and losing even one can be like a shot to the heart? Because our goal is to cover the ground ASAP, preferably in the first year and definitely by summer of the second year. Place your plants based on size at the time of installation; over time, the more robust species and specimens will out compete the lesser ones, and that’s ok as long as the ground stays covered in the future.
Once those plants get going and are competing healthily it’s time to crack open a hard lemonade. You can crack open another one when you start seeing plants move around and find their own way -- which is exactly what we want as they fill in, create layers, and augment the design we kickstarted. You can always thin and transplant -- that’s what gardeners do -- but you’ll be surprised and even thankful at what the plants teach as they shuffle, thrive, and falter. Let that dynamic purpose have its way, especially since you planned for it by using multiple layers and plants suited for the site. You’ve also planned for plants to fill niches -- layers of succession and layers of seasonality, as well as layers in time.
One final strategy to consider when using a matrix is based on how grasses tend to dominate in both a prairie and a garden. In some ways we want lots of grasses, as they are effective at erosion and weed control while providing critical wildlife habitat. Still, we obviously want flowers (and so do pollinators and spiders and birds). Something to consider when choosing forbs is to select plants the spread by rhizome or root runners, as well as those that tend to produce good amounts of seed. That last point is counter intuitive to what we explored above, about not using free-seeding flowers in a small bed; and that still holds true for a small bed. But in larger areas approaching several thousand square feet, we want to have at least a few plants that cast their seed around -- and maybe we especially want those that drop seed near the mother plant to create larger colonies, masses, and drifts.
You can see there’s a lot to consider, but I wouldn’t want to do without at least 40-50% grass cover because of their many benefits to ecosystem function -- either using them in intermingled or matrix designs. If you find grasses starting to tip the balance too much, say 70%, management like early summer mowing, dormant forb overseeding, and definitely planting more plugs in fall will all turn the tide.
*The above is an excerpt from my forthcoming book, Reprairie Suburbia: An Introduction to Natural Garden Design (2022).