Web Map Service (WMS) layers allow organizations to incorporate spatial data from other places, and clients are now routinely integrating WMS layers into their ESRI-centric applications. I'm going to discuss a couple of the cartographic issues that can crop up with this.
When a user switches on a WMS layer, the map viewer requests the data directly from the WMS which returns a map image back to the user’s web browser. The map viewer stacks the various image layers from the different data sources one on top of the other. Normally this would result in only the top image being visible to the user, but the map images can take advantage of the concept of transparency; that is, areas of the map image that contain no data can be made transparent, allowing the images lower in the stack to show through.
While the image stacking method mentioned above works is simple enough, there are situations where it yields unsuitable cartographic results. Annotation can be especially problematic, as a feature’s label may have been placed at exactly the same spot as the label for a different feature from a different WMS data source. Since the WMS layers are raster images, there is no way for different WMS servers to perform label collision detection and reconciliation, which is done for WMS layers drawn from the same data source. Multiple labels drawn at exactly the same location will be illegible.
In addition to the annotation problem, there is a potentially more important issue whereby polygons (district boundaries, for example) that are symbolized with a solid fill color are completely opaque to WMS layers from data sources further down on the image stack. A common internet cartography technique is to symbolize a polygon with a partially transparent fill color, allowing features underneath the polygon to show through. For example, a floodplain feature may be symbolized with a partial transparent fill, allowing all of the land features to be seen through the feature. The end result is a tinge of color across the floodplain area with little effect on the visibility of the coincident features. However, output map images compiled from different sources don't provide transparency options that are well suited for web environments.
So… when integrating WMS layers from various sources into your mapping application, it is important to consider how they will appear to end users. WMS layers can be a convenient way to enrich an ArcIMS/ArcGIS Server application, but are of diminished value if cartographic representation issues detract from the effectiveness of the final map.