To elucidate the water-conducting pathways in living trees by the dye injection method, suitable sample preparation procedures are needed. We evaluated quantitatively the properties and concentrations of three dyes (acid fuchsin, basic fuchsin and safranin) widely used for this purpose, and determined the optimal conditions required to avoid artifacts after dye injection into the sap stream of Pieris japonica D. Don. Among the dyes tested, an aqueous solution of acid fuchsin at a concentration of 0.1% or more was the most useful for delineating water movement. In non-transpiring stem segments, the vertical movement of acid fuchsin by capillarity and diffusion from the dye injection site was limited. However, acid fuchsin moved rapidly in the horizontal direction by capillarity and diffusion, and most xylem cells were stained within 2 h. A delay of more than 2 h between dye injection and examination of the tissues greatly reduces the precision of the method. Use of the dye injection method without appropriate, well-defined experimental procedures may give rise to misleading information about the functional water-conducting pathway in living trees.

The interspecific variation of dye ascent in the stems of 44 broadleaved deciduous species growing in Japan was studied using freeze-dried samples after dye injection. The dye ascending pattern differed both within and between ring-porous and diffuse-porous species. In large earlywood vessels of all ring-porous species, the dye ascended only in the outermost annual ring, and the inner annual rings had lost their water transport function. The dye ascending pattern within the inner annual rings in the ring-porous species was categorized into three types: i) the dye ascended both in the many latewood vessels throughout the latewood and small earlywood vessels; ii) the dye ascended in the many vessels throughout the latewood; and iii) the dye ascended mainly in the late latewood vessels. In diffuse-porous species, the dye ascending pattern within the annual rings also was categorized into three types: i) the dye ascended throughout the annual rings; ii) the dye ascended mainly in the earlywood vessels; and iii) the dye ascended mainly in the latewood vessels. Xylem water distribution was also examined by cryo-SEM in three ring-porous and three diffuse-porous species that had different dye ascending patterns. The water distribution pattern within annual rings was correlated with the dye ascending pattern except for one diffuseporous species (Salix gracilistyla). In this case, water was distributed in the whole region of the annual rings although dye was mainly distributed in the earlywood. These results showed that the functional area of water transport within annual rings differed among ring-porous species and diffuse-porous species.