diff --git a/test_rclcpp/CMakeLists.txt b/test_rclcpp/CMakeLists.txt
index 7120e14e..904cf719 100644
--- a/test_rclcpp/CMakeLists.txt
+++ b/test_rclcpp/CMakeLists.txt
@@ -254,6 +254,26 @@ if(AMENT_ENABLE_TESTING)
         "rclcpp")
     endif()
 
+    # test_multithreaded
+    ament_add_gtest(
+      gtest_multithreaded__${middleware_impl}
+      "test/test_multithreaded.cpp"
+      TIMEOUT 15
+    )
+    if(TARGET gtest_multithreaded__${middleware_impl})
+      target_link_libraries(gtest_multithreaded__${middleware_impl}
+        ${_AMENT_EXPORT_ABSOLUTE_LIBRARIES}
+        ${_AMENT_EXPORT_LIBRARY_TARGETS})
+      target_compile_definitions(gtest_multithreaded__${middleware_impl}
+        PUBLIC "RMW_IMPLEMENTATION=${middleware_impl}")
+      add_dependencies(gtest_multithreaded__${middleware_impl} ${PROJECT_NAME})
+      rosidl_target_interfaces(gtest_multithreaded__${middleware_impl}
+        ${PROJECT_NAME} ${typesupport_impl})
+      ament_target_dependencies(gtest_multithreaded__${middleware_impl}
+        "${middleware_impl}"
+        "rclcpp")
+    endif()
+
     # test_local_parameters
     ament_add_gtest(
       gtest_local_parameters__${middleware_impl}
diff --git a/test_rclcpp/test/test_multithreaded.cpp b/test_rclcpp/test/test_multithreaded.cpp
new file mode 100644
index 00000000..361632ad
--- /dev/null
+++ b/test_rclcpp/test/test_multithreaded.cpp
@@ -0,0 +1,289 @@
+// Copyright 2015 Open Source Robotics Foundation, Inc.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0 //
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include <string>
+#include <utility>
+#include <vector>
+
+#include "gtest/gtest.h"
+
+#include "rclcpp/rclcpp.hpp"
+#include "rclcpp/executors.hpp"
+
+#include "test_rclcpp/msg/u_int32.hpp"
+#include "test_rclcpp/srv/add_two_ints.hpp"
+
+#ifdef RMW_IMPLEMENTATION
+# define CLASSNAME_(NAME, SUFFIX) NAME ## __ ## SUFFIX
+# define CLASSNAME(NAME, SUFFIX) CLASSNAME_(NAME, SUFFIX)
+#else
+# define CLASSNAME(NAME, SUFFIX) NAME
+#endif
+
+static inline void multi_consumer_pub_sub_test(bool intra_process)
+{
+  std::string node_topic_name = "multi_consumer";
+  if (intra_process) {
+    node_topic_name += "_intra_process";
+  }
+
+  auto node = rclcpp::Node::make_shared(node_topic_name, intra_process);
+  auto callback_group = node->create_callback_group(
+    rclcpp::callback_group::CallbackGroupType::Reentrant);
+  auto pub = node->create_publisher<test_rclcpp::msg::UInt32>(node_topic_name, 16);
+
+  std::vector<rclcpp::Subscription<test_rclcpp::msg::UInt32>::SharedPtr> subscriptions;
+  std::atomic_uint counter;
+  counter = 0;
+
+  auto callback =
+    [&counter, &intra_process](test_rclcpp::msg::UInt32::ConstSharedPtr msg,
+      const rmw_message_info_t & info) -> void
+    {
+      ++counter;
+      printf("callback() %4u with message data %u\n", counter.load(), msg->data);
+      ASSERT_EQ(intra_process, info.from_intra_process);
+    };
+
+  rclcpp::executors::MultiThreadedExecutor executor;
+  // Try to saturate the MultithreadedExecutor's thread pool with subscriptions
+  for (uint32_t i = 0; i < executor.get_number_of_threads(); i++) {
+    auto sub = node->create_subscription<test_rclcpp::msg::UInt32>(node_topic_name, 16, callback,
+        callback_group);
+    subscriptions.push_back(sub);
+  }
+
+  executor.add_node(node);
+  auto msg = std::make_shared<test_rclcpp::msg::UInt32>();
+
+  // wait a moment for everything to initialize
+  // TODO(jacquelinekay): fix nondeterministic startup behavior
+  rclcpp::utilities::sleep_for(1_ms);
+
+  // sanity check that no callbacks have fired
+  EXPECT_EQ(0, counter.load());
+
+  ++msg->data;
+  pub->publish(msg);
+
+  // test spin_some
+  // Expectation: The message was published and all subscriptions fired the callback.
+  // Use spin_once to block until published message triggers an event
+  executor.spin_some();
+  EXPECT_EQ(counter.load(), subscriptions.size());
+
+  // Expectation: no further messages were received.
+  executor.spin_once(std::chrono::milliseconds(0));
+  EXPECT_EQ(counter.load(), subscriptions.size());
+
+  // reset counter
+  counter = 0;
+  msg->data = 0;
+
+  auto publish_callback = [&msg, &pub, &executor](rclcpp::timer::TimerBase & timer) -> void
+    {
+      ++msg->data;
+      if (msg->data > 5) {
+        timer.cancel();
+        // wait for the last callback to fire before cancelling
+        rclcpp::utilities::sleep_for(2_ms);
+        executor.cancel();
+        return;
+      }
+      pub->publish(msg);
+    };
+  auto timer = node->create_wall_timer(std::chrono::milliseconds(1), publish_callback);
+
+  executor.spin();
+  EXPECT_EQ(counter.load(), 5 * subscriptions.size());
+}
+
+TEST(CLASSNAME(test_multithreaded, RMW_IMPLEMENTATION), multi_consumer_single_producer) {
+  // multiple subscriptions, single publisher
+  multi_consumer_pub_sub_test(false);
+}
+
+TEST(CLASSNAME(test_multithreaded, RMW_IMPLEMENTATION), multi_consumer_intra_process) {
+  // multiple subscriptions, single publisher, intra-process
+  multi_consumer_pub_sub_test(true);
+}
+
+TEST(CLASSNAME(test_multithreaded, RMW_IMPLEMENTATION), multi_consumer_clients) {
+  // multiple clients, single server
+  auto node = rclcpp::Node::make_shared("multi_consumer_clients");
+  rclcpp::executors::MultiThreadedExecutor executor;
+
+  std::atomic_uint counter;
+  counter = 0;
+  auto callback = [&counter](const std::shared_ptr<test_rclcpp::srv::AddTwoInts::Request> request,
+      std::shared_ptr<test_rclcpp::srv::AddTwoInts::Response> response)
+    {
+      printf("Called service callback: %u\n", counter.load());
+      ++counter;
+      response->sum = request->a + request->b;
+    };
+
+  rmw_qos_profile_t qos_profile = rmw_qos_profile_services_default;
+  qos_profile.depth = executor.get_number_of_threads() * 2;
+  auto callback_group = node->create_callback_group(
+    rclcpp::callback_group::CallbackGroupType::Reentrant);
+  auto service = node->create_service<test_rclcpp::srv::AddTwoInts>(
+    "multi_consumer_clients", callback, qos_profile, callback_group);
+
+  using ClientRequestPair =
+      std::pair<rclcpp::client::Client<test_rclcpp::srv::AddTwoInts>::SharedPtr,
+      test_rclcpp::srv::AddTwoInts::Request::SharedPtr>;
+  using SharedFuture = rclcpp::client::Client<test_rclcpp::srv::AddTwoInts>::SharedFuture;
+
+
+  std::vector<ClientRequestPair> client_request_pairs;
+  for (uint32_t i = 0; i < 2 * executor.get_number_of_threads(); ++i) {
+    auto client = node->create_client<test_rclcpp::srv::AddTwoInts>(
+      "multi_consumer_clients", qos_profile, callback_group);
+    auto request = std::make_shared<test_rclcpp::srv::AddTwoInts::Request>();
+    request->a = i;
+    request->b = i + 1;
+    client_request_pairs.push_back(ClientRequestPair(client, request));
+  }
+
+  executor.add_node(node);
+  rclcpp::utilities::sleep_for(5_ms);
+
+  executor.spin_once();
+  // No callbacks should have fired
+  EXPECT_EQ(0, counter.load());
+
+  {
+    std::vector<SharedFuture> results;
+    // Send all the requests
+    for (auto & pair : client_request_pairs) {
+      results.push_back(pair.first->async_send_request(pair.second));
+    }
+    // Wait on the future produced by the first request
+    auto result = executor.spin_until_future_complete(results.back());
+
+    ASSERT_EQ(result, rclcpp::executor::FutureReturnCode::SUCCESS);
+
+    // Check the status of all futures
+    for (uint32_t i = 0; i < results.size(); i++) {
+      ASSERT_EQ(std::future_status::ready, results[i].wait_for(std::chrono::seconds(0)));
+      EXPECT_EQ(results[i].get()->sum, 2 * i + 1);
+    }
+
+    EXPECT_EQ(counter.load(), client_request_pairs.size());
+  }
+
+  // Reset the counter and try again with spin
+  counter = 0;
+  {
+    std::vector<SharedFuture> results;
+    // Send all the requests again
+    for (auto & pair : client_request_pairs) {
+      results.push_back(pair.first->async_send_request(pair.second));
+    }
+    auto timer_callback = [&executor, &results]() {
+        for (auto & result : results) {
+          if (result.wait_for(std::chrono::seconds(0)) != std::future_status::ready) {
+            return;
+          }
+        }
+        executor.cancel();
+      };
+    auto timer = node->create_wall_timer(std::chrono::milliseconds(3), timer_callback);
+
+    executor.spin();
+
+    // Check the status of all futures
+    for (uint32_t i = 0; i < results.size(); i++) {
+      ASSERT_EQ(std::future_status::ready, results[i].wait_for(std::chrono::seconds(0)));
+      EXPECT_EQ(results[i].get()->sum, 2 * i + 1);
+    }
+    EXPECT_EQ(counter.load(), client_request_pairs.size());
+  }
+}
+
+static inline void multi_access_publisher(bool intra_process)
+{
+  // Try to access the same publisher simultaneously
+  auto context = std::make_shared<rclcpp::context::Context>();
+  std::string node_topic_name = "multi_access_publisher";
+  if (intra_process) {
+    node_topic_name += "_intra_process";
+  }
+
+  auto node = rclcpp::Node::make_shared(node_topic_name, context, intra_process);
+  auto timer_callback_group = node->create_callback_group(
+    rclcpp::callback_group::CallbackGroupType::Reentrant);
+  auto sub_callback_group = node->create_callback_group(
+    rclcpp::callback_group::CallbackGroupType::Reentrant);
+
+  rclcpp::executors::MultiThreadedExecutor executor;
+
+  auto pub = node->create_publisher<test_rclcpp::msg::UInt32>(node_topic_name);
+  // callback groups?
+  auto msg = std::make_shared<test_rclcpp::msg::UInt32>();
+  // use atomic
+  std::atomic_uint timer_counter;
+  timer_counter = 0;
+
+  const size_t iterations = 5 * executor.get_number_of_threads();
+  auto timer_callback =
+    [&executor, &pub, &msg, &timer_counter, &iterations](rclcpp::timer::TimerBase & timer)
+    {
+      if (timer_counter.load() >= iterations) {
+        timer.cancel();
+        std::this_thread::sleep_for(std::chrono::milliseconds(3));
+        executor.cancel();
+        return;
+      }
+      msg->data = ++timer_counter;
+      printf("Publishing message %u\n", timer_counter.load());
+      pub->publish(msg);
+    };
+  std::vector<rclcpp::timer::TimerBase::SharedPtr> timers;
+  // timers will fire simultaneously in each thread
+  for (uint32_t i = 0; i < executor.get_number_of_threads(); i++) {
+    timers.push_back(node->create_wall_timer(std::chrono::milliseconds(1), timer_callback));
+  }
+  std::atomic_uint subscription_counter;
+  subscription_counter = 0;
+  auto sub_callback = [&subscription_counter](const test_rclcpp::msg::UInt32::SharedPtr)
+    {
+      ++subscription_counter;
+      printf("Subscription callback %u\n", subscription_counter.load());
+    };
+  auto sub = node->create_subscription<test_rclcpp::msg::UInt32>(node_topic_name, sub_callback,
+      rmw_qos_profile_default,
+      sub_callback_group);
+  executor.add_node(node);
+  executor.spin();
+  ASSERT_EQ(timer_counter.load(), iterations);
+  ASSERT_EQ(timer_counter.load(), subscription_counter);
+}
+
+TEST(CLASSNAME(test_multithreaded, RMW_IMPLEMENTATION), multi_access_publisher) {
+  multi_access_publisher(false);
+}
+
+TEST(CLASSNAME(test_multithreaded, RMW_IMPLEMENTATION), multi_access_publisher_intra_process) {
+  multi_access_publisher(true);
+}
+
+
+int main(int argc, char ** argv)
+{
+  // NOTE: use custom main to ensure that rclcpp::init is called only once
+  rclcpp::init(argc, argv);
+  ::testing::InitGoogleTest(&argc, argv);
+  return RUN_ALL_TESTS();
+}