diff --git a/jpa/java8/README.md b/jpa/java8/README.md
index 4d762d6f..ddfb9758 100644
--- a/jpa/java8/README.md
+++ b/jpa/java8/README.md
@@ -15,9 +15,27 @@ interface CustomerRepository extends Repository<Customer, Long> {
 }
 ```
 
-* `CustomerRepository.findOne(Long)` effectively "overrides" the default `CrudRepository.findOne(…)` to return `Optional.empty()` instead of `null` in case no uniqe element satisfying the query can be found.
+* `CustomerRepository.findOne(Long)` effectively "overrides" the default `CrudRepository.findOne(…)` to return `Optional.empty()` instead of `null` in case no unique element satisfying the query can be found.
 * `CustomerRepository.findByLastname(…)` does the same for a normal query method.
 
 ## Support for JDK 8's date/time types in the auditing sub-system
 
-* `Customer` extends `AbstractEntity` which contains fields of JDK 8's new date/time API and those can be populated the same way legacy types like `Date` and `Calendar` can.
\ No newline at end of file
+* `Customer` extends `AbstractEntity` which contains fields of JDK 8's new date/time API and those can be populated the same way legacy types like `Date` and `Calendar` can.
+
+## Support for JDK 8' `Stream` in repository methods
+
+JPA repositories can now use `Stream` as return type for query methods to trigger streamed execution of the query. This will cause Spring Data JPA to use persistnce provider specific API to traverse the query result one-by-one.
+
+```java
+interface CustomerRepository extends Repository<Customer, Long> {
+
+  @Query("select c from Customer c")
+  Stream<Customer> streamAllCustomers();
+}
+
+try (Stream<Customer> customers = repository.streamAllCustomers()) {
+  // use the stream here
+}
+```
+
+Note how the returned `Stream` has to be used in a try-with-resources clause as the underlying resources have to be closed once we finished iterating over the result.
\ No newline at end of file