diff --git a/diagrams/join-anti.png b/diagrams/join-anti.png
new file mode 100644
index 0000000..0d30477
Binary files /dev/null and b/diagrams/join-anti.png differ
diff --git a/diagrams/join-inner.png b/diagrams/join-inner.png
index 4a18e8f..18e996d 100644
Binary files a/diagrams/join-inner.png and b/diagrams/join-inner.png differ
diff --git a/diagrams/join-many-to-many.png b/diagrams/join-many-to-many.png
index b00deaf..c8bb150 100644
Binary files a/diagrams/join-many-to-many.png and b/diagrams/join-many-to-many.png differ
diff --git a/diagrams/join-one-to-many.png b/diagrams/join-one-to-many.png
index 8d9bd44..6ddab8b 100644
Binary files a/diagrams/join-one-to-many.png and b/diagrams/join-one-to-many.png differ
diff --git a/diagrams/join-outer.png b/diagrams/join-outer.png
index fc9c361..946a696 100644
Binary files a/diagrams/join-outer.png and b/diagrams/join-outer.png differ
diff --git a/diagrams/join-semi-many.png b/diagrams/join-semi-many.png
new file mode 100644
index 0000000..5ddd109
Binary files /dev/null and b/diagrams/join-semi-many.png differ
diff --git a/diagrams/join-semi.png b/diagrams/join-semi.png
new file mode 100644
index 0000000..ded6df1
Binary files /dev/null and b/diagrams/join-semi.png differ
diff --git a/diagrams/join.graffle b/diagrams/join.graffle
index e2924ce..4159281 100644
Binary files a/diagrams/join.graffle and b/diagrams/join.graffle differ
diff --git a/relational-data.Rmd b/relational-data.Rmd
index 2a7fc6d..31a5cee 100644
--- a/relational-data.Rmd
+++ b/relational-data.Rmd
@@ -388,7 +388,25 @@ Instead you can use a semi-join, which connects the two tables like a mutating j
 flights %>% semi_join(top_dest)
 ```
 
-The inverse of a semi-join is an anti-join. An anti-join keeps the rows that _don't_ have a match, and are useful for diagnosing join mismatches. For example, when connecting `flights` and `planes`, you might be interested to know that there are many `flights` that don't have a match in `planes`:
+Graphically, a semi-join looks like this:
+
+```{r, echo = FALSE, out.width = "50%"}
+knitr::include_graphics("diagrams/join-semi.png")
+```
+
+Only the existence of a match is important; it doesn't match what observation is matched. This means that filtering joins never duplicate rows like mutating joins do:
+
+```{r, echo = FALSE, out.width = "50%"}
+knitr::include_graphics("diagrams/join-semi-many.png")
+```
+
+The inverse of a semi-join is an anti-join. An anti-join keeps the rows that _don't_ have a match:
+
+```{r, echo = FALSE, out.width = "50%"}
+knitr::include_graphics("diagrams/join-anti.png")
+```
+
+Anti-joins are are useful for diagnosing join mismatches. For example, when connecting `flights` and `planes`, you might be interested to know that there are many `flights` that don't have a match in `planes`:
 
 ```{r}
 flights %>%