I collected seeds of the parasite Castilleja, and two of its hosts, Lupinus and Eriophyllum, on 30 August 1993.
On 13 April 1995, I transplanted single Castilleja individuals into large pulp pots (height = 25 cm, width at top = 25 cm) with a randomly assigned pair of hosts: either 2 Lupinus, 2 Eriophyllum, or one individual of each of these host species (mixed hosts, hereafter).
Sample sizes varied because some Castilleja were too small to use: two Eriophyllum hosts, n = 28; mixed hosts, n = 30; and two Lupinus hosts, n = 25.
I randomly subsampled Castilleja individuals for these analyses: for Castilleja grown with two Eriophyllum, n - 26; with mixed hosts, n = 31; and with 2 Lupinus, n = 25.
13% N (mean [+ or -] 1 SEM), while Eriophyllum hosts contained 1.
Qualitative detection of alkaloids verified that all Castilleja grown with two Lupinus hosts contained alkaloids, while all grown with two Eriophyllum did not.
Overall, analysis of alkaloid content revealed four distinct groups of Castilleja: those with (1) two Lupinus hosts, (2) two Eriophyllum hosts, (3) mixed hosts available but no alkaloids in the parasite, and (4) mixed hosts and alkaloids in the parasite.
However, aphid colony growth was equally poor whether the Castilleja that they fed upon was grown with two Eriophyllum or mixed hosts (Tukey post-hoc test: P = 0.
In a previous study of this system, aphids experienced improved survival and produced more offspring when they fed on Castilleja attacking one Lupinus host compared to those attacking one Eriophyllum host (Marvier 1996a).
This could be due to an active preference by Castilleja for Eriophyllum hosts, or because Eriophyllum roots may simply be easier for Castilleja to penetrate than those of Lupinus, especially given that parasitic plants frequently have difficulty penetrating woody host tissues (Kuijt 1969).
Castilleja that had mixed hosts available, but did not contain alkaloids, did not differ from Castilleja grown with two Eriophyllum hosts in terms of nitrogen, flowers, or mass [ILLUSTRATION FOR FIGURE 1 OMITTED].
Although there is no direct chemical evidence that Castilleja were also attached to the Eriophyllum host, it is quite likely that most of these parasites were simultaneously attacking both host species.
24), there was no difference between the two Eriophyllum and the mixed host treatments (Tukey posthoc test: mixed hosts = 2 Eriophyllum hosts [greater than] 2 Lupinus hosts).
Specifically, alkaloid-containing Castilleja with mixed hosts that contained alkaloids were more likely to have been attacking two host plants than those with mixed hosts and no alkaloids or those in either the two Lupinus or two Eriophyllum treatments (see Results: Refinements).
Overall, Castilleja that were grown with both Lupinus and Eriophyllum, and that contained alkaloids, were very likely to be simultaneously attacking both host species.