I know that he spoke of what he knew, and with what chances of observation I have had, I long ago came to the conclusion that the elements of what is commonly called a Dutch garden do not differ so greatly from, those that went to the making of the oldest English herb and flower garden. This being so, when I asked myself how I should lay out a foreground that should be congenial with the picture seen through the window of the marble-panelled room, I knew that the garden should be as like as possible that which would be planted by the porter's wife when the Castle was at its best. The porter's lodge would join on to the gate, and one side of the gateway touches my ground, where the lodge would be; so that, with suggestions from the Chatelaine, who had seen the world, and the Chaplain, who may have been familiar with the earliest gardens in England--the monastery gardens--she would lay out the little bit of ground pretty much, I think, as I have done. In those days people had not get into the way of differentiating between gardens and gardens--there was no talk about “false notes” in design, men did not sleep uneasily o' nights lest they had made an irremediable mistake in giving hospitality to a crimson peony in a formal bed or in failing to dig up an annual that had somehow found a place in a herbaceous border. But a garden bounded by walls must be neat or nothing, and so the porter's wife made a Dutch garden without being aware of what she was doing, and I followed her example, after the lapse of a few hundred years, knowing quite well what I was doing in acting on the principle that the surroundings should suggest the garden. I know now, however, that because William the Conqueror had a fine growth of what we call _Dianthus Caryophylla_ at his Castle of Falaise, we should have scrupulously followed his example. However, the elements of a Dutch garden are geometrical, and within four walls and with four right angles one cannot but be geometrical. One cannot have the charming disorderliness of a meadow bounded by two meandering streams. That is why I know I was right in refusing to allow any irregularity in my treatment of the ground. I put my sundial exactly in the middle and made it the centre for four small beds crossed by a narrow grass path; and except for the simple central design there is no attempt at colour effect. But every one of the little beds is brilliant with tulips or pansies or antirrhinum or wallflowers, as the season suggests. There is the scent of lavender from four clumps--one at each angle of the walls--and over the western coping a pink rose climbs. To be consistent I should confine the growth of this rose to an espalier against the wall. I mean to be consistent some day in this matter and others nearly as important, and I have been so meaning for the past ten years.