Last weekend, I was in Seattle and had the opportunity to visit the Washington Park Arboretum, the Center for Urban Horticulture (both now administered as the University of Washington Botanic Gardens), and the Volunteer Park Conservatory. It was my first visit to the latter, and though it is quite small, I can see why it was recommended to me by some colleagues. That said, I spent most of my time at the arboretum, as it gave me the best opportunity to retrain myself on some of the technical aspects of photography I've let slip (though this is not on display with these photographs).
Shrubby species and cultivars of Cornus, or dogwoods, are often grown ornamentally for their colourful stems. The best results are achieved when plants are pruned back in early spring; the Missouri Botanical Garden's profile on 'Midwinter Fire' bloodtwig dogwood explains two of the typical pruning options for Cornus (cut everything back to 30cm / 12" or cut back a fraction of the oldest stems each year). The Royal Horticultural Society suggests the latter method, in order to maintain the framework in their profile for this cultivar: Cornus sanguinea 'Midwinter Fire'. Both links contain additional details about the ecology and growing requirements for this particular cultivar. For these keenly interested in shrubby dogwoods for their garden, a broader assessment was published in American Nurseryman magazine, Comparing Cornus (PDF). This assessment was based on evaluations at Longwood Gardens.
The question of why the stems colour has only recently begun to be addressed. A query was made about this last year on the UBC BG Forums: Cornus colour and carotenoid pigments, but unfortunately it wasn't answered with this 2010 paper: Gould, KS et al. 2010. Why some stems are red: cauline anthocyanins shield photosystem II against high light stress. J. Exp. Bot. 61(10):2707-2717. doi:10.1093/jxb/erq106.
Note first of all that it is anthocyanins that are responsible (at least in part) for pigmentation in Cornus, not carotenoids. Secondly, the term "high light stress" is a reference to the evidence that shows that plants can receive too much light. There are a number of detrimental physiological reactions to too much light, but among the most easily explainable is the excess absorption of energy by photosynthetic molecules. This results in the production of ROS, or reactive oxygen species (species being used here as a class of molecules, not in the taxonomic sense). The Plant Biology Stress Lab at Manchester University explains ROS: "These are highly reactive derivatives of oxygen that are capable of reacting with and destroying a wide range of biomolecules including DNA, proteins and lipids".
Gould et al. observed that the high concentrations of anthocyanins in red stems helped these stems to be better protected against excess light levels than plants with green stems, also noting "the redder the stems, the greater was the photoprotective advantage as compared with green stems". This observation was made in 5 of the 6 taxa they experimented with, including a shrubby dogwood species, Cornus stolonifera. The outlier was a cultivar of Lobelia erinus, of which the authors wrote: "[this Lobelia was] bred specifically for a dark purple/blue flower, and its exceptionally high levels of anthocyanins (and chlorophylls) in the stems are the likely outcome of artificial selection by plant breeders rather than of physiological requirement. Nonetheless, the inclusion of Lobelia stems in our study serves to indicate that the photoprotective hypothesis is not universally applicable...There remains much to be learned about anthocyanin function in stems".