This appeal requires definition of the statutory term “ordinary high water mark” (Civ. Code, § 830), and application of the rule to a situation in which the land itself moves in greater degree, but with less frequency, than do the tides.

Defendant owned a sandspit bounded on one side by the Pacific Ocean, and on the opposite side by Bolinas Lagoon. On each of these sides, defendant owned the upland and the state the submerged lands. Following state survey, defendant brought a quiet title action. Decree was entered in 1950. Defendant then subdivided the land and sold some lots. It drove a line of iron rails, several feet apart, into the sand at a right angle to the sea, and posted “no trespassing” signs. The state contended that the fence extended into state property and constituted a public nuisance. It brought this action for injunction. Evidence showed that the mean high tide could be established by visual reference to a monument or bench mark of the U. S. Coast and Geodetic Survey. There was also evidence that the sand builds up seaward during the summer months, thus increasing the beach area, and that a reverse movement decreases the land area in winter months. This movement of the sand can account for variations of as much as 80 feet in the point at which the water, at ordinary high tide, touches the shore line. There is room to infer that this variation occurs annually.

The court enjoined defendant from interfering with the public use of the land lying “seaward of the ordinary high water mark,” but declined to fix the location of that mark. Both findings and judgment term the boundary “the ordinary high water mark of the Pacific Ocean as it may fluctuate naturally from time to time.” There are findings that defendant’s fence “extends or may extend from time to time seaward of ordinary high water mark” and that “from time to time, the ordinary high water mark . . . migrates landward due to natural erosion and' seaward due to natural accretion.” Defendant appeals.

We cannot accept defendant’s argument that the 1950 decree established the then state survey line as a fixed boundary. The description of that decree carries the line only to “the ordinary high-water mark on the shore of the Pacific Ocean,” and thence along the shore by courses and distances. Its description of the ocean boundary makes no reference to the state survey. This omission is emphasized by the decree’s description of the lagoon boundary. There the intersecting line runs to the “low water mark in Bolinas Lagoon, as fixed by the . . .State survey.” This contrast compels the conclusion that the 1950 decree leaves open the location of the seaward boundary.

The present judgment, however, is too uncertain to be enforced. Some of this uncertainty stems from the trial court’s view that “ordinary high water mark” is a fluctuating, variable or migrating line.

The misconception is readily understandable. The statutory use of the word “mark” connotes a physical mark of the water upon the ground, and that in turn implies a new mark by each higher wave, with only the highest wave of a particular tide leaving a mark visible for even a briefly appreciable period.

But the true rule is to the contrary. The “high water mark” is not “a physical mark made upon the ground by the waters; it means the line of high water as determined by the course of the tides” (Borax, Consolidated v. City of Los Angeles, 296 U.S. 10, 22 [ 56 S.Ct. 23 , 80 L.Ed. 9 ]).

As it fixes a boundary, “tide” must be viewed as a surface or plane of seawater (Swarzwald v. Cooley, 39 Cal.App.2d 306, 313 [ 103 P.2d 580 ]). This plane at “ordinary high tide” is an average height of the high waters at a particular place over a long period of time (Borax, Consolidated v. City of Los Angeles, supra; Swarzwald v. Cooley, supra). The ebb and flow of the tide, and the varying heights of the several tides, are largely caused by the gravity forces of moon and sun, the former exercising about double the effect of the latter. The varying positions of the two bodies in relation to each other and to the particular point of the earth’s surface being considered, effect substantial differences in the height of the several high tides. The most commonly recognized variations follow the phases of the moon. But the lunar month is not a sufficient period to determine an average of high tides. Rather, the full range of astronomical variants affecting the height of tides is deemed covered only in 18.6 years (Borax, Consolidated v. City of Los Angeles, supra, 296 U.S. 10, 27 ; U.S. Coast and Geodetic Survey, “Tidal Datum Plane, ’ ’ Special Publication No. 135, pp. 76, 81; see also “Relation of the Tide to Property Boundaries,” R. Adm. R. S. Patton, former director, U.S. Coast and Geodetic Survey, 1940, p. 5).

It is apparent that the terms “ordinary high tide,” and “mean high tide,” as used in the cases and statutes, refer to an average over a long period, expressed in terms of height or elevation of the plane of water. Obviously, since it does represent an average, it is a fixed figure. It does not vary, fluctuate or migrate save over a long period of time and even then probably in but minute degree. The plane of average high tide is readily ascertainable by reference to monuments and tidal data of the U.S. Coast and Geodetic Survey. To the extent that the findings, conclusions and judgment imply that “ordinary high tide” is a variable plane, they impart uncertainty.

Of course, reference to ordinary high water as a boundary or “mark” necessarily refers to the point at which the plane of average high tide meets the land (Bolsa Land Co. v. Vaqueros Major Oil Co., 25 Cal.App.2d 75, 80-81 [ 76 P.2d 519 ]; Swarzwald v. Cooley, supra, 39 Cal.App.2d 306, 313 ). To the extent that the land itself moves, as is the case of the sandy beach here in issue, this line does vary.

The evidence, however, does not support the finding that “from time to time” the boundary “migrates landward due to natural erosion and seaward due to natural accretion,” thus constantly transferring ownership of the intermediate land. Although the code provision on the subject of “alluvion” (Civ. Code, §1014) (often termed “accretion”) specifically relates only to land formed upon the bank of a river or stream, it is settled that the like common-law rule applies to tidal shores (Strand Improvement Co. v. Long Beach, 173 Cal. 765, 772-773 [ 161 P. 975 ]; Curtis v. Upton, 175 Cal. 322, 334 [ 165 P. 935 ]; Carpenter v. City of Santa Monica, 63 Cal.App.2d 772, 788 [ 147 P.2d 964 ]). But addition to the land is accretion (and loss of land is deliction), only when the changes are gradual and imperceptible (Carpenter v. City of Santa Monica, supra, p. 789), or “by little and little, by small and imperceptible degrees” (Dana v. Jackson Street Wharf Co., 31 Cal. 118, 120 [89 Am.Dec. 164]). The evidence here is that the beach is some 80 feet wider in summer than in winter. If these changes be constant, in offsetting pairs occurring annually, they can hardly be gradual and imperceptible, and thus cannot meet the definitions of natural accretion and delietion.

No effort was made by the parties to determine whether these movements are of substantially the same distance each year, thus affording a basis for fixing an average, mean, or ordinary line of the shore against which the average plane of the water at high tide may be placed to determine a reasonably definite boundary line. This effort should be made upon retrial. We recognize that the “almost mathematical line bounding Blaekacre” (Rest., Torts, § 943, p. 724) cannot be achieved, but somewhat greater certainty should be possible than the constantly moving line contemplated by the present decree. If not, the trial court should consider whether the uncertainty is so great as to warrant denial of injunctive relief (Rest., Torts, § 943; 27 Cal.Jur.2d 162).

Since retrial must be had, we point to another issue which may be determinable upon additional evidence. The figures of the United States Coast and Geodetic Survey were the only evidence as to height of ordinary high water. As pointed out, these are based on an average of all high tides. The California rule, however, uses the average of all high neap tides (see California decisions cited in Borax, Consolidated v. City of Los Angeles, supra, 296 U.S. 10, 26, fn. 5 ; Carpenter v. City of Santa Monica, supra, 63 Cal.App.2d 772, 784-785 ). Neap tides are those occurring when the moon is in its first and third quarters, and the high neap tides are somewhat lower than the high spring tides occurring at times of new moon and full moon (Borax, Consolidated v. City of Los Angeles, supra, at p. 23; “Tidal Datum Plane,” supra, p. 3; “Relation of the Tide to Property Boundaries,” supra, pp. 3-4). It may well be that available data will correlate the datum here used and the average of all high neap tides. But there is no evidence on that issue, or upon the crucial question of the difference on the ground which would be made by application of the latter figure at this particular point of the coast line. This issue should be covered by evidence. If it shows that the difference is negligible, or that it cannot be determined, the federal standard may be the only evidence.

Judgment reversed.

Salsman, J., and Devine, J., concurred.

A petition for a rehearing was denied June 10, 1966, and respondent’s petition for a hearing by the Supreme Court was denied July 6, 1966. Mosk, J., did not participate therein.