112 STUDIES IN SEEDS AND FRUITS 



where both kinds of seeds occur in the same plant, sometimes, but not 

 usually, distinguished also by external characters. It is the variable 

 group that ofters the best materials for study (p. 91). 



(2) The author then gives the list of about 105 species of seeds 

 on which he experimented, arranged in these three groups. Two- 

 fifths of them are leguminous, and of these three-fourths are claimed 

 by the two groups with impermeable seeds, a fact indicating the pre- 

 dominance of impermeable seeds amongst the Legummosae as compared 

 with other orders, and previously established by the researches of 

 Nobbe, Gola, Ewart, and others (p. 93). 



(3) It is shown that impermeable seed-coats, whilst characterising 

 both buoyant and non-buoyant seeds, are indispensable for seeds that 

 are transported bv currents, and that the coral island in mid-ocean 

 would be deprived of many of its most conspicuous plants if their 

 seeds were pervious to water. The author proceeds to point out that 

 as regards the adaptive relations of impermeability it would be wise to 

 cast one's net widely and adopt the standpoint of Professor Ewart that 

 in seed-impermeability we have a general adaptation to soil-conditions 



(p. 97). . , , . 



(4) The lesson of the plants possessing both permeable and im- 

 permeable seeds distinguished by external characters is supplied by 

 Entada polystachya. In this connection it is remarked that the clue 

 to the origin of these two kinds of seeds in the same plants lies in the 

 shrinking and drying of the soft pre-resting seed, a process which is 

 completed in the impermeable seed, but arrested at an earlier stage in 

 the permeable seed. The cause of the check in the shrinking and 

 drying of the seed is determined by variation in the behaviour of the 

 drying pod (p. 98). 



(5) The same indications are supplied by those numerous plants where 

 permeable and impermeable seeds are associated without differing much 

 in their external characters, as illustrated by desalpinia Sappan and 

 Ipomosa tuberosa (p. lOl). 



(6) The difficulties that may attend the development of imperme- 

 ability in seeds are then illustrated in detail from a study of Dioclea 

 reflexa in the forests of Grenada (p. 103). 



(7) The changes in the condition of the seed-coats associated with 

 loss of impermeability are exemplified by Guilandina bondiicella (p. 105). 



(8) Mould or mildew, by attacking the soft coats of the seed in 

 the moist green pod and destroying the cuticle, is regarded as one of 

 the greatest foes to the development of impermeability. That typical 

 hard resting seeds may be largely proof against the attacks of these 

 minute fungi has been shown by the experiments of Nobbe and Gola. 

 But the author observes that it is not there we should look for risks to 

 the seed from this cause, since it is the so-called unripe soft-coated seed 



