In natural environment, water in apoplastic spaces in most plants freezes by reduction of air temperatures to near subzero temperatures. When water freezes in apoplastic spaces, most plant cells respond by extracellular freezing. [Fujikawa et al. 1999]
Ice formation begins somewhere in the plant after a few degrees of supercooling, and ice propagation proceeds through the extracellular spaces. This creates an extracellular vapor pressure deficit, and cell water is drawn from the protoplasm to the extracellular spaces where it freezes. [Burke et al 1976]
The necessary qualifications for effective supercooling are not fully understood, but they include (1) small cell size; (2) little or no intercellular space for nucleation; (3) relatively low water content; (4) absence of internal nucleators; (5) barriers against external nucleators; (6) a dispersion of cells into independently freezing units which allows for supercooling; and (7) the presence of antinucleator substances which oppose the formation of nucleation. [Sakai & Larcher 1987]
Small volumes of water, or large volumes of very pure water, can be cooled considerably below 0¡C before they freeze. The water in biological systems is found in very small volumes, such as the interiors of cells and organelles, and so can supercool substantially... Ice crystals are only stable if they have a certain minimum size: one cannot, for instance, have one molecule of ice in a volume of liquid water. The minimum size of an ice crystal decreases with sub-freezing temperature, but just below freezing it is large. ... Now the chance that a large number of molecules will spontaneously form themselves into such a crystal is small. [Wolfe]
One occasionally encounters the myth that xylem water supercools substantially. This idea probably arose from the fact that it is difficult to freeze fast-moving water if the stem is chilled locally, because the chilled water continuously "escapes". [Tyree & Zimmermann 2002]
They hypothesized that some of the unexplained additional pressure might also have been associated with the expansion of water during the phase change from liquid to solid. [Cavender-Bares]
Substances in nature than can act as heterogeneous nucleators include: (a) ice nucleation-active (INA) bacteria; (b) other biological molecules and structures; and (c) organic and inorganic debris. Nucleators may be on the plant surface (extrinsic) or, in some cases, within the plant (intrinsic; see below). To function, a potential hetero- geneous nucleator must be in contact with water. Consequently, if the plant surface is dry, extrinsic nucleators will be inefective. However, during radiation frosts in many climates, moisture will tend to condense onto plant surfaces so giving an opportunity for any heterogeneous nucleators present on the plant surface to function. Snow and sleet can also initiate freezing in plants. [Pearce]
An apparent misunderstanding seems to stem from a confusion of the two very different kinds of sap, xylem sap and phloem sap.
The confusion would seem to be exacerbated by statements such as the following, from The Encyclopaedia Britannic Online: "The xylem and phloem within the stem distribute the water and sap throughout the plant." Water and sap would seem to be separate entities in the stem, perhaps the meaning was 'water [in xylem] and sap [in phloem] respectively?' Some online biology texts mirror this interpretation.
"Xylem sap consists mainly of water and inorganic ions, although it can contain a number of organic chemicals as well." [Wikipedia]
"Phloem sap is an aqueous solution. Its most common solute is sugar, mostly sucrose. The sucrose concentration can be as high as 30% by weight. Phloem sap also contains minerals, amino acids, and hormones." [Hedger]
All evidence indicates that the xylem sap is the sole source of the crystals that form in crystallofolia; i.e., water in the secondary xylem from the roots or elsewhere.
As the temperature continues to drop, a second phase of release of the heat of fusion of water is detectable.... [Taiz & Zeiger]
When most of the water in a sample is still liquid it has a heat of fusion near that of pure water (79 cal/g). [Burke et al 1976]
The one extra requirement of the process we observed (transition of the entire body of the liquid into solid) is that the liquid must be far enough below its freezing point that the heat of crystallization does not bring it above its freezing point (and for most liquids, including water, that heat is fairly large, so the liquid actually has to be well below its freezing point). [Zaphod]
... solutes from the vessels reach the ray cells via the contact pits. These are extremely large pits between the ray cells and vessels.
Radial intercellular canacules between the ray cells offer a potential rout for apoplastic flow driven by the transpiration.... The canacules probably originate from fusion of schizogeneously-formed intercellular spaces. The apoplastic component of lateral transport may be significant.
The basal driving force for radial flow of water in the direction of the stem periphery, however, may be the high osmolarity of the sieve tube contents.