catch moved toward Japan, effort expanded lon- 

 gitudinally, and effort continued to increase in 

 negative rotation during the study period. An- 

 other example of a trend might be found in the 

 tendency for the velocity of the apparent move- 

 ment to increase during the study period. 



These anomalies may result to some extent 

 from variability in the oceanic environment. 

 The problem of relating the abundance and dis- 

 tribution of fish to changes in the oceanic en- 

 vironment (we use the term oceanic environ- 

 ment in the abstract sense to denote the set of 

 all variables with which the albacore coact or 

 react as well as the variability of each and its 

 correlation with other members of the set) is 

 fraught with difficulties involving the measure- 

 ment of the true abundance and position of the 

 fish. In addition, it is difficult to determine 

 which subset of the many environmental vari- 

 ables is actually most relevant in determining 

 the distribution or abundance of the fish, to 

 develop methods of measuring the appropriate 

 subset of these variables, and finally to develop 

 techniques for the study of the relation between 

 the abundance and distribution of fish and the 

 relevant variables of the environment. 



Some progress has been made in this study 

 toward measuring the apparent abundance and 

 apparent spatial distribution of the albacore, 

 but we have not furthered an understanding of 

 the complex interrelations of the albacore and 

 its environment beyond some of the study on 

 the correlation between apparent abundance of 

 albacore and sea-surface temperature (Johnson, 

 1962). Instead, however, we will mention brief- 

 ly some of the concepts that have given us dif- 

 ficulty in supporting our empirical observations 

 by a cause and effect relationship. For simpli- 

 city we will relate our discussion to tempera- 

 ture but it could relate equally well to any other 

 variable or combination of variables, again not 

 only with respect to the absolute levels of the 

 variables, but also to their variability and their 

 mutual intercorrelations. The problem of mea- 

 suring temperature in a way which is appropri- 

 ate for fishery investigations has been dis- 

 cussed by Bell and Pruter (1958). For the al- 

 bacore problem we would need to estimate the 

 average temperature for each 10° quadrangle of 

 the North Pacific Ocean for some finite time 

 period. But there are at least two, probably 

 omnipresent, sources of bias in estimates of 

 the average temperature of a 10° quadrangle of 

 ocean. The first source of bias arises from 

 the logistic constraints of practical vessel 



tracks that prevent taking a random sample of 

 temperatures in the quadrangles of interest. 

 Even an average temperature based upon a 

 more practical systematic, rather than random, 

 sampling plan--even ignoring the mathematical 

 bias that is sometimes inherent in systematic 

 sampling designs (see Cochran, 1953)--can be 

 severely biased. The extent of the bias would 

 depend on the thermal gradient and whether the 

 cruise track was parallel or normal to the iso- 

 thermal lines. It is in fact rather intriguing to 

 consider that some features depicted in ocean- 

 ographic atlases may be charted in a position 

 that is dislocated owing to the direction of the 

 cruise track. The second source of bias in- 

 volves determining the locations of those tem- 

 peratures that are within the cube of water that 

 are actually relevant to the behavior of the 

 fish. In some instances surface temperatures 

 may be relevant; in others, temperature at 

 some constant depth, or possibly even temper- 

 atures on a surface of constant dissolved oxy- 

 gen or density. 



Another difficulty in interpreting the relation 

 between the distribution and abundance of fish 

 and the oceanic environment is that a fish's 

 response to the environment is time-dependent 

 and not time-independent as is usually implicit- 

 ly assumed when relating ocean temperatures 

 to fish distribution. The time-dependence is a 

 compensatory phenomenon in which an animal's 

 lethal temperature or temperature preferen- 

 dum, for example, is a function of its past 

 thermal history. Pioneering studies of this 

 "physiological memory" were accomplished by 

 Fry (1947). Over the subsequent two decades 

 there have been many studies of the compensa- 

 tory process. Most of these studies have been 

 of acclimation, which refers to compensatory 

 changes that organisms undergo in response to 

 changing levels of an environmental variable 

 under simplified laboratory conditions, but not 

 of acclimatization which refers to the compen- 

 satory process under the more complex, and of 

 course more realistic, natural situation (see, 

 e.g.. Hoar, 1966: p. 296). Although acclimati- 

 zation has not been studied for the albacore it 

 almost certainly operates and influences ther- 

 mal and other preferenda of the albacore. While 

 acclimatization may produce small changes in 

 temperature preferendum, these small changes 

 could involve rather large horizontal distances. 

 Although the nature of the acclimatory period 

 with respect to intensity and duration of expo- 

 sure of an organism to various levels is not 



31 



