1890.] Training Walls in an Estuary like the Mersey. 143 



model of the Mersey estuary, from near Warrington to the open sea 

 beyond the bar, made to a horizontal scale of 30 ^ 00 , and a vertical 

 scale of -^-^0, and with a bed formed of fine Bagshot sand. The expe- 

 riments were directed to the solntion of two problems, namely, 

 (1) The influence of training walls in the wide npper estuary on the 

 channel below Liverpool, and across the bar ; and (2) The effects of 

 training works in the lower estuary on the channel across the 

 bar. 



The model was first worked, without modification, till a fair repro- 

 duction was obtained of the existing conditions of the estuary. 

 Training walls, made of strips of tin, were then inserted in the 

 model, following the lines of the Manchester Ship- Canal Scheme of 

 1884, down the middle of the upper estuary, for which the present 

 line of canal, in course of construction, skirting the Cheshire shore, 

 was substituted in 1885. This modification soon produced a change 

 in the model, which no previous working had effected ; for, though 

 the channel between the training walls was deepened, the upper 

 estuary began to silt up as the working of the model proceeded, and 

 the channel immediately below Liverpool began to shoal, till at last 

 the main navigation channel below the " narrows " became very 

 shallow for some distance. This result solves a very much disputed 

 question as to the effects of these training works, and, together with 

 the results of the Seine model, affords grounds for the conclusion 

 that training walls placed in a sandy estuary, where sandbanks exist 

 below them, will lead to accretion behind them. As the channel in 

 the lower estuary of the Mersey is almost wholly maintained by the 

 tidal water flowing into and out of the estuary above, accretion in the 

 upper estuary would necessarily produce a deterioration in the 

 channel below. 



The model was next restored to its original form, and training- 

 walls were inserted in the lower estuary, in continuation of the 

 narrow channel between Liverpool and Birkenhead, gradually 

 diverging out with a trumpet-shaped outlet, so as not to impede the 

 tidal influx into the upper estuary. The scour through this trained 

 channel gradually washed away a sandbank to the north of the out- 

 let, and eventually formed a channel to the north of the* previous 

 channel, towards Formby Point, in the model, with a minimum 

 depth considerably more than was previously obtained across the 

 bar. 



The bed of the model was then restored to its previous state ; but 

 the training walls in the lower estuary were retained, and a sand- 

 bank impeding the outlet was removed to the level of the bar. The 

 scour through the trained channel, with this arrangement, produced 

 a more direct and uniform channel than in the previous case, with a 

 similar increase in depth. 



