The correct answer is that it only increases the spread. Note that we are only multiplying the values. The number of values, or their counts, are unchanged. Multiplying a set of normally distributed numbers by a value effectively changes the standard deviation of the distribution they were drawn from. Now why is this important to know? Remember that randn generates values with standard deviation one, whereas the images we've been using are of type uint8 and range from zero to 255. 

What do you think happens when you add the results of randn directly to an image? Let's find out. Time to use a new image. If you look carefully you'll be able to see three moons and a shadow. Now we generate our noise image and add it to the original. Not really very different, is it? This is because the values that randn generated are really small compared to the image. Let's scale up the values. Now you can see the noise affecting the image. How about we increase this further? And more. Now it's really hard to see the moons, isn't it?