Often it is useful to plot two or more data sets or population on the same axes.  This makes it easy to compare the data, using differing colors to show multiple populations at the same time.

Overlays may be created on either univariate (right) or bivariate displays (see bottom). Within the layout editor you can edit the color and order of the items, to customize the look of the overlay graph that best highlights your data.  Once an overlay has been defined, the layout editor can create the same graph for many different samples or sets of samples, by iterating over a group in the workspace.

To create an overlay, drag a graph into the Layout Editor.  Then drag a second on top of the first one in the layout.  As the mouse moves into the original graph item, you'll see that it highlights its borders to signify that it will accept the contents of the drag.   Drop the second graph (or more, if you had multiple nodes selected in the workspace) on top of the original, and an overlay is automatically created.

Histograms can optionally be offset when they are overlaid, so as to distinguish the curves better. The feature, as well as how much of an offset is used are specified in the Annotate pane of the Graph Info dialog, summoned by double clicking the graph (or selecting it and choosing Get Info... from the Layout menu).

The Overlay Legend

As soon as you release the second node on the first, the graph will change into an overlay.   Overlays are recognizable by the existence of a legend on the right side of the graph.  The number of the rows in the legend shows the number of layers in the overlay.  The graphs often change from black and white to color, as each data set gets its own color within the overlay.

The legend provides the user interface to edit several aspects of the overlay.  Click the mouse once on the legend to select it.  Then pass the mouse over different areas of the legend to see to change the cursor to signify the different editing operation that are available.

When the mouse passes over an edge of the legend, the cursor will become a hand. Clicking the mouse on the outside edge of the legend supports moving or resizing the legend. 

If you move the mouse of the text description of the layers, you'll see the cursor become a bar with vertical arrows sticking out of it.  This tools allows you to reorder the legend, which can be a very important step as the topmost layers are drawn on top of the ones below them in the legend.  Because events can obscure others, you may want to have the smaller population on top of the larger ones, though this depends greatly on the context of your analysis.  Dragging the order of the layers will cause the graph to be redrawn, so you can quickly see which layers you want to feature in the top positions.  If you hold down the option key, you'll see that the cursor changes from a vertical reordering tool into a trash can.  This signifies that option-clicking a layer in the legend of a graph will delete that layer from the graph.

Clicking the mouse in the colored box at the left side of the legend will pop up a "rainbow palette," to let you change the colors of the layers in the graph.  If the graph is a univariate (a histogram or cumulative distribution function) the line weight and fill style are also editable from a popup menu under the line style box (just to the right of the color box, and only on univariate graphs).

The legend is used to control all of the attributes of the graphs appearance. By default the legend is visible if there are multiple items in an overlay, and hidden for single histograms or 2D plots. If you want to edit the color, line weight, or dashing of a single graph, you can double click on the graph, and set the Show Legend attribute of the Annotation panel.