diff --git a/R/summary-slide.R b/R/summary-slide.R
index 553b969..a6710d6 100644
--- a/R/summary-slide.R
+++ b/R/summary-slide.R
@@ -59,7 +59,7 @@
 #' `slide_dbl(x, sum)`. This is extremely slow with large window sizes and
 #' wastes a lot of effort recomputing nearly the same information on each
 #' window. It can be made slightly faster by moving the sum to C to avoid
-#' intermediate allocations, but it still fairly slow.
+#' intermediate allocations, but it is still fairly slow.
 #'
 #' A second alternative is to use an _online_ algorithm, which uses information
 #' from the previous window to compute the next window. These are extremely
diff --git a/man/summary-slide.Rd b/man/summary-slide.Rd
index 2621988..2f9a981 100644
--- a/man/summary-slide.Rd
+++ b/man/summary-slide.Rd
@@ -158,7 +158,7 @@ function on each full window. This is what is done by using, for example,
 \code{slide_dbl(x, sum)}. This is extremely slow with large window sizes and
 wastes a lot of effort recomputing nearly the same information on each
 window. It can be made slightly faster by moving the sum to C to avoid
-intermediate allocations, but it still fairly slow.
+intermediate allocations, but it is still fairly slow.
 
 A second alternative is to use an \emph{online} algorithm, which uses information
 from the previous window to compute the next window. These are extremely