We have found few pollen grains of buckwheat in deposits containing early iron age pottery fragments. Redeposition or vertical movement of pollen is probable but there is no clear proof so we couldn't omit any alternative.
In the Netherlands, there is quite some discussion about the first occurrence of Fagopyrum. Written sources fail to go back earlier than 1392, but there is one sample with one single carbonized buckwheat fruit dating 550-700 AD from Dommelen, published by Pals 1988 (in Dutch: Pals, J.P., 1988. Akkerbouw in het Middeleeuwse Dommelen, in: Phyto-archeologische studies. Unpublished PhD-thesis UvA, Amsterdam.). In this site, however, also later medieval habitation traces have been excavated, so intrusion of younger material may be problematic here. I have a publication in prep with three colleagues about all early buckwheat finds from the Netherlands and Belgium, both archaeobotanical and historical data.
However, there are somewhat more regular finds of pollen from mainly Karolingian times. Henk van Haaster extensively deals with the buckwheat finds in his thesis (also in Dutch: Van Haaster, H., 2008. Archeobotanica uit ’s-Hertogenbosch. Milieuomstandigheden, bewoningsgeschiedenis en economische ontwikkelingen in en rond een (post)middeleeuwse groeistad, Groningen Archaeological Studies 6. Barkhuis Publishing, Eelde). He is convinced of an early Medieval cultivation of buckwheat, although all pollen is identified as Fagopyrum spec (including F. tataricum). Prehistoric finds of buckwheat are absent from the Netherlands, and maybe F. tataricum may even be more likely in such old cases in eastern Europe!
Just an addition to my previous answer: we have now submitted our manuscript. We discovered that in the past there was a flaw in the search function of the European Pollen Database, and I had, thus, overlooked many sites. The total amount of records from western Eurasia as far as I know of is around 235.
Dear Zdenek: I now know of eight records from the Czech Republic, of which one also dates to the Iron Age.
Dear Otto: I found over 30 sites from the Netherlands and Belgium, does this correspond with the amount of records in your inventarisation? There are indeed some records from prehistoric times.
I cannot give exact numbers, as we have not considered pollen records (at least not yet), and I know these by far outnumber records of macroremains before 1400 AD!
May I assume that most of your hits for the Netherlands and Belgium do concern pollen?