ON THE FOOD OF THE MENHADEN. 115 



If now these be telescoped together into their natural position in the sides of the 

 mouth cavity of the fish, then A will be the most anterior ; B, 0, I), E will follow in 

 a closely arranged series; the branchial filaments m, n, o, p, each overlapping its next 

 posterior neighbor, will be next the operculum which covers them. In like manner 

 the gill-rakers a, b, c, d, each underlapping its next anterior neighbor, will cover the 

 the interior surface of the mouth cavity, and this in a very perfect manner. The gill- 

 rakers of A all project straight forward (a) ; those of B (b) upon the upper section of 

 the gill-arch curve upward, as in fact the remaining members of the series all do 

 (/>, c, d,) in order to reach up to the upper section of each preceding gill-arch. 



Reference to the diagram (fig. 1) of the gill-arches in place will show that they 

 project downwards into the roof of the mouth, each one hanging lower down than the 

 one preceding it; and the curving upward of the gill-rakers upon the upper section 

 of each of the three arches (2, 3, 4) makes such an arrangement as to completely cover 

 the roof of the mouth with these curving gill-rakers from the protruding gill arches. 

 Upon the upper section of gill-arch 3 is indeed a double row of gill-rakers ; the regular 

 series of upward curving ones — the continuation of c projecting from the upper 

 segment of this arch — to which is added another one (a?) projecting downward and 

 backward to meet y of the next posterior series. (Part of the series c is omitted in 

 the drawing in order to expose series x to view.) The lower edge of x therefore fits 

 into the edge of series y, which curves abruptly upward to meet it. The rudimentary 

 gill-arch 5 carries a short and stiff row of gill-rakers only. If the mouth of a men- 

 haden be opened the observer is confronted by a throat cavity completely lined with 

 a layer of these beautifully adjusted lamelke of gill -rakers, overlapping each other 

 in the most perfect manner — as I have attempted to explain in the foregoing ; their 

 function, namely, that of extracting food from the water, is as perfectly performed 

 as the sequel will, it is hoped, aid in showing. 



The water passing into the mouth of the fish will pass through this system of 

 gill-rakers and make its exit posteriorly from under the opercula. The water passing 

 through the gill-slit between m and n will be filtered by the series of gill-rakers b; 

 the water passing through the gill-slit between n and o will be filtered by c; the 

 water passing through the gill-slit between o and p will be filtered by d, y, x; in each 

 case, of course, the upper portion of the gill-arches with their filaments will be bathed 

 by a large part of the water as it passes through. That the column of water in so passing 

 through this apparatus maybe deprived of the organisms living in it is made evident 

 by the finer structure of the gill-rakers themselves, as illustrated in the other figures 

 of plate 2. They constitute rows of thin elastic bony blades arranged upon the 

 anterior edge of each gill-arch, as before described; fig. 3 represents the basal part 

 of six of the gill-rakers attached to their support; their full length is not, however, 

 given. Each individual blade — covered with a delicate epithelium, in which are 

 mucus and large-branched pigment cells — is provided with two rows of hooked barbs, 

 one row upon either of its sides, which so project in between from all of the adjoining 

 gill-rakers as to guard each interspace with two rows of these barbs. A diagram of 

 these gill-rakers in cross-section is given in fig. 5, showing the attachment of the barbs 

 to the exterior (toward the operculum) edge of each gill-raker blade, allowing them to 

 project inwards toward the current of water passing through the mouth cavity of 

 the fish. Finally, in much higher magnification, 178 diameters, are given, fig. 4, 



