• annual grass;
  • fire;
  • great basin;
  • invasion;
  • life stage;
  • seeding


1. Seeding is an important management tool in aridland restoration, but seeded species often fail to establish. Previous research has largely focused on the technical aspects of seeding with little effort directed at identifying demographic processes driving recruitment failures.

2. In tilled plots, in each of 3 years, we estimated life stage transition probabilities for three species commonly used in sage steppe restoration. We also took similar measurements on seed sown by managers following four major fires.

3. Point estimates and associated Bayesian confidence intervals demonstrated germination probabilities that were consistently high, averaging 0·72. However, estimates suggest only 17 and 7% of the germinated seeds emerged in the tilled plots and fire sites, respectively. Following emergence, survival across the seedling, juvenile and adult transitions averaged 0·72. This suggests the transition from a germinated seed to an emerged seedling was the major bottleneck in recruitment. Although most individuals died during emergence, this was not always the principal source of variation in recruitment across sites.

4.Synthesis and applications. Processes occurring after emergence, such as mortality during spring and summer drought, may contribute to site-to-site variation in recruitment but are unlikely to be the main causes of restoration failures. Instead, recruitment may largely be determined by processes occurring during emergence, such as freezing and thawing of the seedbed, development of physical soil crusts and pathogen attack on germinated seeds. Using tools such as seed coatings and soil amendments to manage processes inhibiting emergence and developing seed mixes with higher emergence probabilities are likely to greatly improve restoration outcomes in the sage steppe and similar aridland systems.