2006 Annual Report 4d.Progress report. This report serves to document research conducted under a reimbursable agreement between ARS and the University of California, Davis. Additional details of research can be found in the report for the parent project 5358-21000-034-00D,Production Systems to Promote Yield and Quality of Grapes in the Pacific Northwest. Dr. Julie Tarara and collaborators conducted the following research towards the agreements objectives: Continued development and testing of the 'Trellis Tension Monitor' was successful in 2005 and continues through 2006. The concept is simple: grapevines are not self supporting, and as they grow during the season their increasing mass imposes increasing weight on the trellis that supports them. A load cell inserted inline with the 'cordon' wire (main horizontal support wire of the trellis) continuously detects the increasing tension in the trellis wire that is generated by the growing vines and fruit. The seasonal pattern of increase in tension can be used to monitor vine growth, fruit development, and ultimately to estimate final yield. We continued comparison of 'open-ended' and 'closed system' trellis tension monitors. In the 'open' system, the load cell is allowed to sense increasing weight on the trellis as far along the row as possible, whereas in the 'closed' system we have circumscribed the load cell's sensing distance by isolating it within the short span of cordon wire between two trellis posts. We term the 'closed' system the 'Bridge' because a tubular steel brace spans the post-to-post distance and allows the trellis wire from the remainder of the row to bypass the load cell's detection zone. The 2005 dataset was the first full season of quality data in 10 commercial juice grape vineyards. At each cooperator's vineyard an 'open' load cell system was paired with a 'Bridge' system for direct comparison. A complete data set from 2005 was the necessary precursor for our first attempt at an empirically-based yield estimation protocol from the 2006 trellis tension measurements in these same vineyards. In the research block of juice grapes, because we had a full quality dataset in 2004, we made our first test of a simple yield estimation protocol using the 2005 tension measurements. Our approach was empirical and mimicked the juice processors' generic prediction protocol that is applied to traditional, static hand samples. We compared the ratio of cordon wire tension-to-yield for the previous year (2004) against the increase in tension during 2005. Such an empirical approach has advantages in its simplicity, but its limitations are location specificity and an assumption of similarity between the years compared, especially in the ratio between shoot growth and fruit growth). In a research block of wine grapes, five pairs of 'Bridge' and 'open' trellis tension monitoring systems were adapted to the 'modified sprawl' training system that is common to wine grapes in our area. The early-season tension data were peculiar, a response that we hope is associated with the mechanics of our having modified the trellis and installed the load cells. The systems appear to have "settled" based on the current 2006 tension data that demonstrate an expected growth curve up to this point. All fruit associated with the trellis tension monitors was hand harvested and weighed, giving us yield data directly associated with the experimental vines. In the commercial vineyards, the cooperating juice processors shared their records of yield for the entire blocks that contained our trellis tension monitors. Cursory analysis indicates that our empirically-derived yield estimates from single trellis monitors can be within 10-25 % of the actual harvest values for the block. Our efforts during the spring of 2006 focused on servicing the trellis tension monitors across all installations. Data processing routines were streamlined to help with turn-around time between downloading and archiving/analysis during the season. Currently, we monitor 59 trellis tension monitors, the sum of which produce approximately 39,648 tension values per week. Data from on-site loggers are downloaded twice each week. Ancillary work continues with bi-weekly allometric measurements of vine growth paired with smaller destructive samples to estimate shoot, leaf, and fruit mass across the season. After 2006, all data will be synthesized, statistical analyses conducted, and the results submitted for publication to horticultural and specialized viticultural journals. In 2005 and continuing in 2006 we are collaborating with the Viticulture Extension Specialist at Washington State University-Prosser to develop a web page (http://www.fruit.wsu.edu/Grapeweb/cropload.htm) for dissemination of the near-continuous tension data (twice-weekly updates). During 2006 the website also will post yield estimates for our two research vineyards and the ten commercial sites.