In an effort to identify possible alternate hosts for the grapevine leafroll- associated viruses and vitiviruses, numerous Vitis and non-Vitis species were surveyed within and surrounding nine vineyards in Napa County, CA, with vines showing typical leafroll disease symptoms. Three riparian areas in Napa and Yolo counties that were not near vineyards but that included a number of endemic Vitis spp. with red leaves were also identified.

Outside the vineyard, Vitis spp. and the most common woody non-Vitis spp. that were present within a 25 m2 area adjacent to symptomatic V. vinifera blocks were collected. Within each vineyard, at least three vines from a block adjacent to non-cultivated Vitis and/or non-Vitis were collected.

Herbaceous plants were collected by establishing three 1 m2 plots located within 10 meters of the symptomatic V. vinifera collected the previous Fall. Within each plot, three replicate samples of all the dicotyledonous species were collected and pooled. Like species were also pooled between plots due to the large number of different species.

VIRUS DETECTION: Vitis and woody samples collected in Fall 2008 were assayed for GLRaV-1, -2, -3, -4, -5, -9, GVA, GVB, and GVD, by both conventional and real-time RT-PCR. Only real-time RT-PCR was used to detect these same viruses in herbaceous samples collected in Spring 2009, and Vitis and non-Vitis samples collected in Fall 2009 and 2010. Real-time RT-PCR assays were run in duplicate or triplicate, and data was analyzed quantitatively by measuring the cycle threshold (CT).

VIRAL DNA SEQUENCES: Total RNA from the samples positive by real-time RT-PCR was re-amplified, also using RT-PCR. Nucleic acid sequences were compared using BLAST from the National Center for Biotechnology Information, and aligned and edited.

VITIS GENOTYPE: The non-cultivated Vitis samples from outside the vineyards were analyzed at 18 Simple Sequence Repeat (SSR) DNA markers. The resulting DNA profiles were compared to over 1,200 unique grape DNA profiles. To identify the species background of the samples, a genetic approach was used to calculate the probability of observing each unique profile from this study and those of five diverse V. vinifera cultivars, based on gene frequencies from a population of 200 random V. vinifera

cultivars. The results were expressed as the inverse of these probabilities divided by the average of the probabilities of the five V. vinifera profiles.

DNA fingerprinting indicated 48 unique multilocus genotypes. Four of these were known scion or rootstock cultivars and were excluded from further analyses.