Fixed bug: The option on the Range tab of nonlinear regression "Don't fit points when... " doesn't work if X is defined as a series.

Background 1: The X column on an XY data table can be defined in two ways.

• Most commonly, you enter actual values into each cell. Or you can use the Insert Series command (it has its own button and is on the Insert menu) to insert a set of X values. When you do this, the generated values are not shown in bold, and you can edit them just like values you entered. 
• Alternatively, you can define the X column to be a series. You specify the first X value and the increment (in the Format Data Table dialog), and Prism creates an infinite list of X values. This series is shown in bold because you can't edit the values. 

Background 2: The Range tab of nonlinear regression has options that let you tell Prism to not fit points if X is less than (or greater than) some value you enter. 

The fixed bug: That feature (ignoring points with certain X values) only works in the usual situation where you entered X values. If the X values are defined as a series, so appear bold in the data table, this feature simply doesn't work. The choices appear on the Range tab, but Prism ignores them and fits all the data. 

Three alternative work arounds:

• Go back to the data tables and exclude or delete the values you don't want to fit, and you'll get the fit you expected. 
• Change the table to have a normal X column and enter values (or use the Create Series button which creates individual X values, which is different than defining the whole column to be a series). The Create Series command works well and is verstatile, so really can replace the older idea of defining the X column to be a series. The word "series" is used for both approaches, which are distinct. Create Series enters a set of X values that you can edit. Defining the X column to be a series creates an infinite list of bold X values that you cannot edit (so are shown in bold). 
• Write a user-defined equation that fits one model when X has one range of values, and another model when X is outside that range. This page shows you how.

Starting with Prism 6.04 and 6.0e, Prism presentx a warning when you try to restrict the range while the X values are defined as a series.