317 
data accordingly afford opportunity to trace the cyclic move- 
ment in production in this period. 
The July pulse has a duration of 32 days,—from June 28 to 
July 30,—with a maximum amplitude of 8.16 em.’ per m.* on 
the 14th. Its mean falls on the 20th. This was a month of 
falling river levels with slight interruptions by local rains, of 
rising temperature and of falling nitrates, but of increasing 
sewage contamination (Pl. XLIV.) as shown by the rising chlo- 
rine and oxygen consumed. The pulse presents a very sud- 
den drop in production from 6.40¢m.‘ on the 16th to .92 em.’ on 
the 21st, followed by an immediate recovery to 6.91 on the 
23d. I amataloss fora satisfactory explanation of this fluc- 
tuation. There is no change in levels at Havana (see page 
160) which suggests flood, though there is a slight increase in 
turbidity (Table III.) and was a rise of .1 ft. at Copperas 
Creek on the 20th which does not appear in the Havana gage 
readings. The chemical analysis of the sample taken on the 
21st (Table X.and Pl. XLIV.) contains evidence of some dis- 
turbance in conditions. There isa sharp decline in nitrates, 
nitrites, and oxygen consumed, with a check in the rising 
chlorine, while free and albuminoid ammonia and total or- 
ganic nitrogen move upwards. Had the oxygen consumed 
risen and the free ammonia fallen, all indices would point. to- 
ward access of recent storm water carrying silt into the stream 
and locally diluting the plankton, though not materially af- 
fecting the hydrograph. In any event the fluctuation in pro- 
duction is correlated with a localized disturbance in chemical 
conditions suggesting in some particulars restricted access of 
recent storm waters. 
The average production for this month (4.69) is higher 
than in any other year excepting 1894 and 1895, due it seems 
to the somewhat stable conditions of continued decline in ley- 
els, with slight overflow sufficiently prolonged (3 weeks at 6 ft.) 
to afford time for breeding plankton in the waters of overflow, 
though apparently similar floods in 1896 resulted in much light- 
er production (Pl. X.). The main difference in hydrographic 
